The opinion of the court was delivered by

Clapp, S. J. A. D.

Defendant appeals from a county district court judgment taken against him for breach of contract. Contract and breach are admitted or assumed on appeal; the only issue is damages.

Under the contract plaintiff leased to defendant an automatic phonograph, a juke box, for two years and agreed to supply records and replace parts wearing out. Proceeds of the operation were to be shared on a specified basis, but with a minimum of $20 per week to be paid plaintiff by defendant. Defendant, it is claimed, repudiated the contract; and plaintiff never installed the machine.

The court gave plaintiff judgment for $836 — that is, the sum of $20 per week for two years, less apparently the costs *130plaintiff -would have been put to, had he performed the contract, less also depreciation on the machine.

Defendant makes two points. The first rests on plaintiff’s testimony that the component parts of the very machine he had intended to lease defendant were, after the breach, rented to others. Defendant argues that the amount plaintiff thus realized should have been credited on the claim sued upon.

Defendant would have us apply here the rule obtaining on the breach of an agreement to lease realty; that is, he claims the measure of the lessor’s damages here is the difference between the agreed rental and the rental value of the property. His contention further is that even though under the agreement before us, the lessor is obliged to perform some personal services, he, in order to establish the rental value, has the burden of proving what he received on a reletting. But see McDermott v. DeMeridor Co., 80 N. J. L. 67 (Sup. Ct. 1910 — the point was not passed upon on the appeal, 82 N. J. L. 530 (E. & A. 1911)) referring to Milage v. Woodward, 186 N. Y. 252, 78 N. E. 873 (Ct. App. 1906) where a canal boat, captain-owner, crew and horses were hired; note 17 A. L. R. 2d 968, 972, 990. All this raises questions we need not consider.

Plaintiff, passing the questions (or most of them), meets the argument by referring to his testimony, not contradicted, that:

“The equipment called for by this agreement was readily available in the market. But locations were very hard to get.”

We think the position plaintiff takes on the matter is sound. Where, as here, a plaintiff lessor agrees to lease an article of which the supply in the market is for practical purposes not limited, then the law would be depriving him of the benefit of his bargain if on the breach of the agreement, it required his claim against the lessee to be reduced by the amount he actually did or reasonably could realize on a reletting of the article. Eor if there had been no breach and another customer had appeared, the lessor could as well have secured another such article and entered into a second lease. In case of the *131breach of the first lease, he should have the benefit of both bargains or not — in a situation whore the profit on both would be the same — be limited to the profit on the second of them.

An illustration with figures may make this more graphic. If the agreed rental under the lease amounts to $2,040, the cost of installation and of furnishing records and parts to $500, and the depreciation on the juke box over the period of the lease to $700, the lessor stands to make $840 on the deal. If another customer presents himself, the lessor will buy another juke box, which he is entitled to enter on his books at cost and depreciate in the same way as he does with the first. Thus, if he makes the same agreement with the second customer, he will make another $840 on the second lease. If the first lessee repudiates his agreement, the purchase of an additional machine will, of course, be unnecessary, because the first machine can be leased to the second customer. In such a situation, under defendant’s theory, the lessor would receive as damages for this repudiation only the $2,040 rental agreed on under the first lease, less the $2,040 rental for the same machine under the second lease, or nothing. This would leave the lessee only the $840 profit he will make under the second lease; whereas had the first lessee lived up to his bargain, the lessor’s profits would have been $840 on each of two leases, or $1,680.

We conclude that the proper measure of damages here is the difference between the contract price and the cost of performing the first contract, as the court apparently held below. In the case of realty which (unlike the juke box) is specific and not to be duplicated on the market, the lessor could not properly lease it to another for the same period unless the first lease were broken or terminated. In such a case the lessor should not be awarded two profits merely because of the first lessee’s default.

So in general we may say that gains made by a lessor on a lease entered into after the breach are not to be deducted from his damages unless the breach enabled him to make the gains. The recoverable damages in the case of a contract are *132such as may reasonably be within the contemplation of the parties at the time of the contract, Patco Products, Inc., v. Wilson, 5 N. J. 543, 547 (1950); and with that in view, we should not in the present case deny lessor the benefit of his bargain. Contrast Zeliff v. Sabatino, 15 N. J. 70 (1954), referring to the rule obtaining in ease of a fraud.

Restatement of Contracts § 336(c) and Illustrations 6 and 7, and 5 Corbin, Contracts, § 1041 (1951) support these propositions. Cf. 3 Williston, Sales (rev. ed. 1948), § 583a. The principles, however, seem not to be widely recognized, but there are cases dealing with various sorts of contracts, which can be said to sustain them. Sullivan v. McMillan, 37 Fla. 134, 19 So. 340 (Sup. Ct. 1896); Olds v. Mapes-Reeves Const. Co., 177 Mass. 41, 58 N. E. 478 (Sup. Jud. Ct. 1900); Mt. Pleasant Stable Co. v. Steinberg, 238 Mass. 567, 131 N. E. 295, 15 A. L. R. 749 (Sup. Jud. Ct. 1921); J. E. Rishel Furniture Co. v. Stuyvesant Co., 123 Misc. 208, 204 N. Y. S. 659 (Mun. Ct. 1924); Allen v. Murray, 87 Wis. 41, 57 N. W. 979, 982 (Sup. Ct. 1894); cf. J. A. Laporte Corporation v. Pennsylvania-Dixie Cement Corp., 164 Md. 642, 165 A. 195, 168 A. 844, 108 A. L. R. 1474 (Ct. App. 1933); Tradesman Co. v. Superior Mfg. Co., 147 Mich. 702, 111 N. W. 343, 112 N. W. 708 (Sup. Ct. 1907); Barron G. Collier, Inc., v. Kindy, 146 Minn. 279, 178 N. W. 584 (Sup. Ct. 1920); Ware Bros. Co. v. Cortland Cart & Carriage Co., 210 N. 7. 122, 103 N. E. 890 (Ct. App. 1913); DeMoss v. Beryllium Corp., 358 Pa. 470, 58 A. 2d 70, 72 (Sup. Ct. 1948).

These principles lead us logically into questions with which we have no concern here — in particular, the question as to a seller’s remedy on a sale of goods where he (like the plaintiff here) has for practical purposes the capacity to supply all probable customers — a matter as to which the law may perhaps be governed by statute, R. S. 46 :30-70, or by precedent, 3 Williston, Sales (rev. ed. 1948), § 583a. But see 5 Corbin, Contracts, § 1100, 1039 (1951). Note, too, the change in the Sales Act proposed by the Uniform Commercial Code — Sales § 2-708 (1952), also Comment 2 *133thereon — a matter apparently overlooked in the 1955 amendment to the section. We limit our opinion to the situation at hand.

Defendant’s second point is that the liquidated damage clause in the lease agreement excluded any recovery by plaintiff. The clause reads:

“In the event of a breach of this agreement by the proprietor [defendant], the said proprietor agrees to pay as liquidated damages and not as a penalty or forfeiture, a sum equal to the average weekly share of the Company [plaintiff] prior to the breach, multiplied by the number of weeks remaining in the unexpired term of the agreement, to be paid immediately at the breach of this agreement.”

This clause seems to provide for liquidated damages in a case where the machine has been in use for some weeks and a base has thus been established for determining plaintiff’s average weekly earnings from the machine. Defendant’s argument is that implicit in the clause is an agreement by the seller to waive damages if a breach occurred (as it did) before the machine was installed. We see in this clause no implication that such was the intention of the parties, and find the argument to be without merit.

Affirmed.

Cabdozo, J.

The plaintiff built a country residence for the defendant at a cost of upwards of $77,000, and now sues to recover a balance of $3,483.46, remaining unpaid. The work of construction ceased in June, 1914, and the defendant then began to occupy the dwelling. There was no complaint of defective performance until March, 1915. One of the specifications for the plumbing work provides that “ all wrought iron pipe must be well, galvanized, lap welded pipe of the grade known as ‘ standard pipe ’ of Reading manufacture.” The defendant learned in March, 1915, that some of the pipe, instead of being made in Reading, was the product of other factories. The plaintiff was accordingly directed by the architect to do the work anew. The plumbing was then encased within the walls except in a few places where it had to be exposed. Obedience to the order meant more than the substitution of other pipe. It meant the demolition at great expense of substantial parts of *241the completed structure. The plaintiff left the work untouched, and asked for a certificate that the final payment war due. Refusal of the certificate was followed by this suit.

The evidence sustains a finding that the omission of the prescribed brand of pipe was neither fraudulent nor willful. It was the result of the oversight and inattention of the plaintiff’s subcontractor. Reading pipe is distinguished from Cohoes pipe and other brands only by the name of the manufacturer stamped upon it at intervals of between six and seven feet. Even the defendant’s architect, though he inspected the pipe upon arrival, failed to notice the discrepancy. The plaintiff tried to show that the brands installed, though made by other manufacturers, were the same in quality, in appearance, in market value and in cost as the brand stated in the contract — that they were, indeed, the same thing, though manufactured in another place. The evidence was excluded, and a verdict directed for the defendant. The Appellate Division reversed, and granted a new trial.

We think the evidence, if admitted, would have supplied some basis for the inference that the defect was insignificant in its relation to the project. The courts never say that one who makes a contract fills the measure of his duty by less than full performance. They do say, however, that an omission, both trivial and innocent, will sometimes be atoned for by allowance of the resulting damage, and will not always be the breach of a condition to be followed by a forfeiture (Spence v. Ham, 163 N. Y. 220; Woodward v. Fuller, 80 N. Y. 312; Glacius v. Black, 67 N. Y. 563, 566; Bowen v. Kimbell, 203 Mass. 364, 370). The distinction is akin to that between dependent and independent promises, or between promises and conditions (Anson on Contracts [Corbin’s ed.], sec. 367; 2 Williston on Contracts, sec. 842). Some promises are so plainly independent that they can never *242by fair construction be conditions of one another. (Rosenthal Paper Co. v. Nat. Folding Box & Paper Co., 226 N. Y. 313; Bogardus v. N. Y. Life Ins. Co., 101 N. Y. 328). Others are so plainly dependent that they must always be conditions. Others, though . dependent and thus conditions when there is departure in point of substance, will be viewed as independent and. collateral when the departure is insignificant (2 Williston on Contracts, secs. 841, 842; Eastern Forge Co. v. Corbin, 182 Mass. 590, 592; Robinson v. Mollett, L. R., 7 Eng. & Ir. App. 802, 814; Miller v. Benjamin, 142 N. Y. 613). Considerations partly of justice and partly of presumable intention are to tell us whether this or that promise shall be placed in one class or in another. The simple and the uniform will call for different remedies from the multifarious and the intricate. The margin of departure within the range of normal expectation upon a sale of common chattels will vary from the margin to be expected upon a contract for the construction of a mansion or a “ skyscraper.” There will be harshness sometimes and oppression in the implication of a condition when the thing upon which labor has been expended is incapable of surrender because united to the land, and equity and reason in the implication of a like condition when the subject-matter, if defective, is in shape to be returned. From the conclusion that promises may not be treated as dependent to the extent of their uttermost minutiae without a sacrifice of justice, the progress is a short one to the conclusion that they may not be so treated without a perversion of intention. Intention not otherwise revealed may be presumed to hold in contemplation the reasonable and probable. If something else is in view, it must not be left to implication. There will be no assumption of a purpose to visit venial faults with oppressive retribution.

Those who think more of symmetry and logic in the development of legal rules than of practical adaptation to the attainment of a just result will be troubled by a classi*243fication where the lines of division are so wavering and blurred. Something, doubtless, may be said on the score of consistency and certainty in favor of a stricter standard. The courts have balanced such considerations against those of equity and fairness, and found the latter to be the weightier. The decisions in this state commit us to the liberal view, which is making its way, nowadays, in jurisdictions slow to welcome it (Dakin & Co. v. Lee, 1916, 1 K. B. 566, 579). Where the line is to be drawn between the important and the trivial cannot be settled by a formula. “ In the nature of the case precise boundaries are impossible ” (2 Williston on Contracts, sec. 841). The same omission may take on one aspect or another according to its setting. Substitution of equivalents may not have the same significance in fields of art on the one side and in those of mere utility on the other. Nowhere will change be tolerated, however, if it is so dominant or pervasive as in any real or substantial measure to frustrate the purpose of the contract (Crouch v. Gutmann, 134 N. Y. 45, 51). There is no general license to install whatever, in the builder’s judgment, may be regarded as “ just as good ” (Easthampton L. & C. Co., Ltd., v. Worthington, 186 N. Y. 407, 412). The question is one of degree, to be answered, if there is doubt, by the triers of the facts (Crouch v. Gutmann; Woodward v. Fuller, supra), and, if the inferences are certain, by the judges of the law (Easthampton L. & C. Co., Ltd., v. Worthington, supra). We must weigh the purpose to be served, the desire to be gratified, the excuse for deviation from the letter, the cruelty of enforced adherence. Then only can we tell whether literal fulfilment is to be implied by law as a condition. This is not to say that the parties are not free by apt and certain words to effectuate a purpose that performance of every term shall be a condition of recovery. That question is not here. This is merely to say that the law will be slow to impute the purpose, in the silence of the parties, where the significance *244of the default is grievously out of proportion to the oppression of the forfeiture. The willful transgressor must accept the penalty of his transgression (Schultze v. Goodstein, 180 N. Y. 248, 251; Desmond-Dunne Co. v. Friedman-Doscher Co., 162 N. Y. 486, 490). For him there is no occasion to mitigate the rigor of implied conditions. The transgressor whose default is unintentional and trivial may hope for mercy if he will offer atonement for his wrong (Spence v. Ham, supra).

In the circumstances of this case, we think the measure of the allowance is not the cost of replacement, which would be great, but the difference in value, which would be either nominal or nothing. Some of the exposed sections might perhaps have been replaced at moderate expense. The defendant did not limit his demand to them, but treated the plumbing as a unit to be corrected from cellar to roof. In point of fact, the plaintiff never reached the stage at which evidence of the extent of the allowance became necessary. The trial court had excluded evidence that the defect was unsubstantial, and in view of that ruling there was no occasion for the plaintiff to go farther with an offer of proof. We think, however, that the offer, if it had been made, would not of necessity have been defective because directed to difference in value. It is true that in most cases the cost of replacement is the measure (Spence v. Ham, supra). The owner is entitled to the money which will permit him to complete, unless the cost of completion is grossly and unfairly out of proportion to the good to be attained. When that is true, the measure is the difference in value. Specifications call, let us say, for a foundation built of granite quarried in Vermont. On the completion of the building, the owner learns that through the blunder of a subcontractor part of the foundation has been built of granite of the same quality quarried in New Hampshire. The measure of allowance is not the cost of reconstruction. “ There may be *245omissions of that which could not afterwards be supplied exactly as called for by the contract without taking down the building to its foundations, and at the same time the omission may not affect the value of the building for use or otherwise, except so slightly as to be hardly appreciable ” (Handy v. Bliss, 204 Mass. 513, 519. Cf. Foeller v. Heintz, 137 Wis. 169, 178; Oberlies v. Bullinger, 132 N. Y. 598, 601; 2 Williston on Contracts, sec. 805, p. 1541). The rule that gives a remedy in cases of substantial performance with compensation for defects of trivial or inappreciable importance, has been developed by the courts as an instrument of justice. The measure of the allowance must be shaped to the same end.

The order should be affirmed, and judgment absolute directed in favor of the plaintiff upon the stipulation, with costs in all courts.

McLaughlin, J. (dissenting).

I dissent. The plaintiff did not perform its contract. Its failure to do so was either intentional or due to gross neglect which, under the uncontradicted facts, amounted to the same thing, nor did it make any proof of the cost of compliance, where compliance was possible.

Under its contract it obligated itself to use in the plumbing only pipe (between 2,000 and 2,500 feet) made by the Reading Manufacturing Company. The first pipe delivered was about 1,000 feet and the plaintiff's superintendent then called the attention of the foreman of the subcontractor, who was doing the plumbing, to the fact that the specifications annexed to the contract required all pipe used in the plumbing to be of the Reading Manufacturing Company. They then examined it for the purpose of ascertaining whether this delivery was of that manufacture and found it was. Thereafter, as pipe was required in the progress of the work, the foreman of the subcontractor would leave word at its *246shop that he wanted a specified number of feet of pipe, without in any way indicating of what manufacture. Pipe would thereafter be delivered and installed in the building, without any examination whatever. Indeed, no examination, so far as appears, was made by the plaintiff, the subcontractor, defendant’s architect, or any one else, of any of the pipe except the first delivery, until after the building had been completed. Plaintiff’s architect then refused to give the certificate of completion, upon which the final payment depended, because all of the pipe used in the plumbing was not of the kind called for by the contract. After such refusal, the subcontractor removed the covering or insulation from about 900 feet of pipe which was exposed in the basement, cellar and attic, and all but 70 feet was found to have been manufactured, not by the Reading Company, but by other manufacturers, some by the Cohoes Rolling Mill Company, some by the National Steel Works, some by the South Chester Tubing Company, and some which bore no manufacturer’s mark -at all. The balance of the pipe had been so installed in the building that an inspection of it could not be' had without demolishing, in part at least, the building itself. - s ■ • •' ■ ' ■ j

I am of the opinion the trial court was right in directing a verdict for the defendant. The plaintiff agreed that all the pipe used should be of the Reading Manufacturing Company. Only about two-fifths of it, so far as appears, was of that kind. If more were used, then the burden of proving that fact was -upon the plaintiff, which it could easily have done, since it knew where the pipe was obtained. The question of substantial performance of a contract of the character of the one under consideration depends in no small degree upon the good faith of the contractor. If the plaintiff had intended to,- and. had complied with the terms of the contract except as: to minor omissions, due to inadvertence, then he might be allowed to recover the contract price, less the amount *247necessary to fully compensate the defendant for damages caused by such omissions. (Woodward v. Fuller, 80 N. Y. 312; Nolan v. Whitney, 88 N. Y. 648.) But that is not this case. It installed between 2,000 and 2,500 feet of pipe, of which only 1,000 feet at most complied with the contract. No explanation was given why pipe called for by the contract was not used, nor was any effort made to show what it would cost to remove the pipe of other manufacturers and install that of the Reading Manufacturing Company. The defendant had a right to contract for what he wanted. He had a right before making payment to get what the contract called for. It is no answer to this suggestion to say that the pipe put in was just as good as that made by the Reading Manufacturing Company, or that the difference in value between such pipe and the pipe made by the Reading Manufacturing Company would be either “ nominal or nothing.” Defendant contracted for pipe made by the Reading Manufacturing Company. What his reason was for requiring this kind, of pipe is of no importance. He wanted that and was entitled to it. It may have been a mere whim on his part, but even so, he had a right to this ldnd of pipe, regardless of whether some other kind, according to the opinion of the contractor or experts, would have been “ just as good, better, or done just as well.” He agreed to pay only upon condition that the pipe installed were made by that company and he ought not to be compelled to pay unless that condition be performed. (Schultze v. Goodstein, 180 N. Y. 248; Spence v. Ham, supra; Steel S. & E. C. Co. v. Stock, 225 N. Y. 173; Van Clief v. Van Vechten, 130 N. Y. 571; Glacius v. Black, 50 N. Y. 145; Smith v. Brady, 17 N. Y. 173, and authorities cited on p. 185.) The rule, therefore, of substantial performance, with damages for unsubstantial omissions, has no application. (Crouch v. Gutmann, 134 N. Y. 45; Spence v. Ham, 163 N. Y. 220.)

*248What was said by this court in Smith v. Brady (supra) is quite applicable here: “ I suppose it will be conceded that everyone has a right to build his house, his cottage or his store after such a model and in such style as shall best accord with his notions of utility or be most agreeable to his fancy. The specifications of the contract become the law between the parties until voluntarily changed. If the owner prefers a plain and simple Doric column, and has so provided in the agreement, the contractor has no right to put in its place the more costly and elegant Corinthian. If the owner, having regard to strength and durability, has contracted for walls of specified materials to be laid in a particular manner, or for a given number of joists and beams, the builder has no right to substitute his own judgment or that of others. Having departed from the agreement, if performance has not been waived by the other party, the law will not allow him to allege that he has made as good a building as the one he engaged to erect. He can demand payment only upon and according to the terms of his contract, and if the condition's on which payment is due have not been performed, then the right to demand it does not exist. To hold a different doctrine would be simply to make another contract, and would be giving to parties an encouragement to violate their engagements, which the just policy of the law does not permit.” (p. 186.)

I am of the opinion the trial court did not err in ruling on the admission of evidence or in directing a verdict for the defendant.

For the foregoing reasons I think the judgment of the Appellate Division should be reversed and the judgment of the Trial'Term affirmed.

Hiscock, Ch. J., Hogan and Crane, JJ., concur with Cardozo, J.; Pound and Andrews, JJ., concur with McLaughlin, J.

Order affirmed, etc.

—Judgment unanimously modified on the law and as modified affirmed without costs and matter remitted to Supreme Court for further proceedings in accordance with the following Memorandum: Defendant appeals from a judgment of Supreme Court awarding plaintiffs damages for defendant’s breach of contract for the construction of an addition to plaintiffs’ home. Defendant in his brief challenges only that aspect of the judgment that awarded damages to plaintiffs for the difference between the market value of the structure had it been completed pursuant to the terms of the contract and the market value of the structure as actually completed. We agree with defendant’s assertion that the record does not support the court’s award for diminution in value, because no such proof was presented.

At trial plaintiffs produced two experts who testified that defendant failed to construct the addition in a good and workmanlike manner. They further testified that the inadequate structural support of the addition rendered unusable the third floor of the addition, which plaintiffs had intended to use as a master bedroom and bathroom. The appeal by defendant, as limited by his brief (see, Ciesinski v Town of Aurora, 202 AD2d 984; Hodge v LoRusso, 181 AD2d 1009), does not contest those findings of fact.

The general rule in cases of faulty construction is that the measure of damages is the market value of the cost to repair the faulty construction (see, American Std. v Schectman, 80 AD2d 318, lv denied 54 NY2d 604). The court erred in applying the "difference in value rule”, as initially set forth by Justice Cardozo in Jacob & Youngs v Kent (230 NY 239, 241), which is limited to instances where the builder’s failure to perform under a construction contract is "both trivial and innocent”, such that damages may be measured by the diminution in value of the building rather than the cost of tearing *937apart the structure and properly completing the project. Where, as here, the defect arising from the breach of the contract "is so substantial as to render the finished building partially unusable and unsafe, the measure of damage is ’the market price of completing or correcting the performance' " (Bellizzi v Huntley Estates, 3 NY2d 112, 115, quoting 5 Williston, Contracts § 1363, at 3825 [rev ed]X Thus, on the facts found by the court, plaintiffs are entitled to the market value of the cost of correcting the deficiencies in the addition arising from defendant’s breach.

The trier of fact is in the best position to evaluate the credibility of the witnesses, who gave conflicting testimony concerning the cost of repair to the addition. Therefore, we modify the judgment appealed from by vacating the court’s award of $10,000 for diminution in value due to inadequate structural support, and we remit the matter to Supreme Court for further findings of fact on the actual cost of repair for inadequate structural support and direct that judgment be entered accordingly. (Appeal from Judgment of Supreme Court, Oswego County, Hurlbutt, J.—Breach of Contract.) Present—Lawton, J. P., Fallon, Wesley, Callahan and Davis, JJ.

Parker, J.

delivered the opinion of the court.

It may be assumed, that the labor performed by the plaintiff, and for which he seeks to recover a compensation in this action, was commenced under a special contract to labor for the defendant the term of one year/for the sum of one hundred and twenty dollars, and that the *486plaintiff has labored but a portion of that time, and has voluntarily failed to complete the entire contract.

It is clear, then, that he is not entitled to recover upon the contract itself, because the service, which was to entitle him to the sum agreed upon, has never been performed.

But. the question arises, can the plaintiff, under these circumstances, recover a reasonable sum lor the service he has actually performed, under the count m quantum meruit.

Upon this, and questions of a similar nature, the decisions to be found in the books are not easily reconciled.

It has been held, upon contracts of this kind for labor to be performed ai a specified price, that the party who voluntarily fails to fulfil the contract by performing the whole labor contracted for, is not entitled to recover any thing for the labor actually performed, however much he may have done towards the performance, and this has been considered the settled rule of law upon this subject.

2 Pick. 267, Stark v. Parker; 2 Mass. 147, Faxon v. Mansfield; 12 Johns. 165, McMillen v. Vanderlip; 13 Johns. 94, Jennings v. Camp; 19 Johns, 337, Reab v. Moor; 8 Cowen, 63, Lantry v. Parks; 9 Barn, & Cres. 92, Sinclair v. Bowles; 2 Stark. Rep. 256, Spain v. Arnott.

That such rule in its operation may be very unequal, not to say unjust, is apparent.

A party who contracts to perform certain specified labor, and who breaks his contract in the first instance, without any attempt to perform it, can only be made liable to pay the damages which the other party has sustained by reaso# of such non performance, which in many instances may be trifling — whereas a party who in good faith has entered upon the performance of his contract, and nearly completed it, and then abandoned the further performance — although the other party has had the full benefit of all that has been done, and has pur-haps sustained no actual damage — is in fact subjected to *487a loss of all which lias been performed, in the nature of damages for the non fulfilment of the remainder, upon the technical rule, that the contract must be fully performed in order to a recovery of any part of the compensation.

By the operation of this rule, then, the party who attempts performance muy be placed in a much worse situation than he who wholly disregards his contract, and the other party may receive much more, by the breach of the contract, than the injury which he has sustained by such breach, and more than he could be entitled to were he seeking to recover damages by an action.

The case before os presents an illustration. Had the plaintiff in this case never entered upon the performance of his contract, the damage could not probably have been greater than some small expense and trouble incurred in procuring another to do the labor which he had contracted to perform. Rut having entered upon the performance, and labored nine and a half months, the value of which labor to the defendant as found by the jury is ⅜95, if the defendant can succeed in this defence,he in fact receives nearly five sixths of the value of a whole year’s labor, by reason of the breach of contract by the plaintiff a sum not only utterly disproportionate to any probable, not to say possible damage which could have resulted from the neglect of the plaintiff to continue the remaining two and an half months, but altogether beyond any damage which could have been recovered by the defendant, liad the plaintiff done nothing towards the fulfilment of his contract.

Another illustration is furnished in Lantry v. Parks, 8 Cowen, 83. There the defendant hired the plaintiff fora year, at ten dollars per month. The plaintiff worked ten and an half months, and then left saying he would work no more for him. This was on Saturday — on Monday the plaintiff returned, and offered to resume his work, but the defendant said he would employ him no longer. *488The court held that the refusal of the defendant on Saturday was a violation of his contract, and that he could recover nothing for the labor performed.

There are other cases, however, in which principles have been adopted leading to a different result.

It is said, that where a party contracts to perform certain work, and to furnish materials, as, for instance, to build a house, and the work is done, but with some variations from the mode prescribed by the contract, yet if the other party has the benefit of the labor and materials he should be bound to pay so much as they are reasonably worth. 2 Stark. Ev. 97, 98; 7 Pick. 181, Hayward v. Leonard; 8 Pick. 178, Smith v. First Cong. Meeting House in Lowell; 4 Cowen, 564, Jewell v. Schroeppel; 7 Green. 78, Hayden v. Madison; Bull. N. P. 139; 4 Bos. & Pul. 355; 10 Johns. 36; 13 Johns. 97; 7 East, 479.

A different doctrine seems to have been holden in Ellis v. Hamlen, 3 Taunt. 52, and it is apparent, in such cases, that if the house has not been built in the manner specified in the contract, the work has not been done. The party has no more performed what he contracted to perform, than he who has contracted to labor for a certain period, and failed to complete the time.

It is in truth virtually conceded in such cases that the work has not been done, for if it had been, the party performing it would be entitled to recover upon the contract itself, which it is held he cannot do.

Those cases are not to be distinguished, in principle, from the present, unless it be in the circumstance, that where the party has contracted to furnish materials, and do certain labor, as to build a house in a specified manner, if it is not done according to the contract, the party for whom it is built may refuse to receive it — elect to take no benefit from what has been performed — and therefore if he does receive, he shall be bound to pay the value— whereas in a contract for labor, merely, from day to day, the party is continually receiving the benefit of the con*489tract under an expectation that it will be fulfilled, and cannot, upon the breach of' it, have an. election to refuse to receive what has been done, and thus discharge himself from payment.

Rut we think this difference in the nature of the contracts docs not justify the application of a different rule in relation to them.

The party who contracts for labor merely, for a certain period, does so with full knowledge that he must, from the nature of the case, be accepting part performance from day to day, if the other party commences the performance, and with knowledge also that the other may eventually fail of completing the entire term.

If under such circumstances he actually receives a benefit from the labor performed, over and above the damage occasioned by the failure to complete, there is as much reason why he should pay the reasonable worth of what has thus been done for his benefit, as there is when he enters and occupies the house which has been built for him, but not according to the stipulations of the contract, and which lie perhaps enters, not because he is satisfied with what has been done, but because circumstances compel him to accept it such as it is, that lie should pay for the value of the house.

Where goods are sold upon a special contract as to their nature, quality, and price, and have been used before their inferiority has been discovered, or other circumstances have occurred which have rendered it impracticable or inconvenient for the vendee to rescind the contract in toto, it seems to have been the practice formerly to allow the vendor to recover the stipulated price, and the vendee recovered by a cross action damages for the breach of the contract. “ But according to the later and more convenient practice, the vendee in such case is allowed, in an action for the price, to give evidence of the inferiority of the goods in reduction of damages, and the plaintiff who has broken his contract is not entitled *490to recover more than the value of the benefits which the defendant has actually derived from the goods ; and where the latter has derived no benefit, the plaintiff'cannot recover at all.” 2 Stark. Ev. 640, 642; 1 Starkie’s Rep. 107, Okell v. Smith.

So where a person contracts for the purchase of a quantity of merchandize, at a certain price, and receives a delivery of part only, and he keeps that part, without any offer of a return, it has been held that he must pay the value of it. 5 Barn. & Cres. Shipton v. Casson; Com. Dig. Action F. Baker v. Sutton; 1 Camp. 55, note.

A different opinion seems to have been entertained, 5 Bos. & Pul. 61, Waddington v. Oliver, and a different decision was had, 2 Stark. Rep. 281, Walker v. Dixon.

There is a close analogy between all these classes of cases, in which such diverse decisions have been made.

If the party who has contracted to receive merchandize, takes a part and uses it, in expectation that the whole will be delivered, which is never done, there seems to be no greater reason that he should pay for what he has received, than there is that the party who has received labor in part, under similar circumstances, should pay the value of what has been done for his benefit.

It is said, that in those cases where the plaintiff has-been permitted to recover there was an acceptance of what had been done. The answer is, that where the contract is to labor from day to day, for a certain period, the party for whom the labor is done in truth stipulates to receive it from day to day, as it is performed, and although the other may not eventually do all he has contracted to do, there has been, necessarily, an acceptance of what has been done in pursuance of the contract, and the party must have understood when he made the contract that there was to he such acceptance.

If then the party stipulates in the outset to receive part performance from time to time, with a knowledge that the whole may not be completed, we see no reason *491why he should not equally he holden to pay for the amount of value received, as where he afterwards takes the benefit of what has been done, with a knowledge that the whole which was contracted for has not been performed.

In neither case has the contract been performed. In neither can an action be sustained on the original contract.

In both the party has assented to receive what is done. The only difference is, that in the one case the assent is prior, with a knowledge that all may not be performed, in the other it is subsequent, with a knowledge that the whole has not been accomplished.

We have no hesitation in holding that the same rule should be applied to both classes of cases, especially, as the operation of the rule will be to make the party who has failed to fulfil his contract, liable to such amount of damages as the other party has sustained, instead of subjecting him to an entire loss for a partial failure, and thus making the amount received in many cases wholly disproportionate to the injury. 1 Saund. 320, c; 2 Stark. Evid. 643.

It is as “ hard upon the plaintiff to preclude him from recovering at all, because he has failed as to part of his entire undertaking,” vvhere his contract is to labor for a certain period, as it can be in any other description of contract, provided the defendant has received a benefit and value from the labor actually performed.

We hold then, that where a party undertakes to pay upon a special contract for the performance of labor, or the furnishing of materials, he is not to be charged upon such special agreement until the money is earned according to the terms of it, and where the parties have made, an express contract the law will not imply and raise a contract different from that which the parties have entered into, except upon some farther transaction between the parties.

*492In case of a failure to perform such special contract, by the default of the party contracting to do the service, if the money is not due by the terms of the special agreement he is not entitled to recover for his labor, or for the materials furnished, unless the other party receives what has been done, or furnished, and upon the whole case derives a benefit from it. 14 Mass. 282, Taft v. Montague; 2 Stark. Ev. 644.

But if, where a contract is made of such a character, a party actually receives labor, or materials, and thereby derives a benefit and advantage, over and above the damage which has resulted from the breach of the contract by the other party, the labor actually done, and the value received, furnish a new consideration, and the law thereupon raises a promise to pay to the extent of the reasonable worth of such excess. This may be considered as making a new case, one not within the original agreement, and the party is entitled to “ recover on his new case, for the work done, not as agreed, but yet accepted by the defendant.” 1 Dane’s Abr. 224.

If on such failure to perform the whole, the nature of the contract be such that the employer can reject what has been done, and refuse to receive any benefit from the part performance, he is entitled so to do, and in such case is not liable to be charged, unless he has before assented to and accepted of what has been done, however much the other party may have done towards the performance. He has in such case received nothing, and having contracted to receive nothing but the entire mat. ter contracted for, he is not bound to pay, because his express promise was only to pay on receiving the whole, and having actually received nothing the law cannot and ought not to raise an implied promise to pay. But where the party receives value — takes and uses the materials, or has advantage from the labor, he is liabe to pay the reasonable worth of what he has received. 1 Camp. 38, Farnsworth v. Garrard. And the rule is the same wheth' *493er it was received and accepted by the assent of the party prior to the breach, under a contract by which, from its nature, he was to receive labor, from time to time until the completion of the whole contract ; or whether it was received and accepted by an assent subsequent to the performance of all which was in fact done. If he received it under such circumstances as precluded him from rejecting it afterwards, that does not alter the case • — it has still been received by his assent.^

In fact we think the technical reasoning, that the performance of the whole labor is a condition precedent, and the right to recover any thing dependent upon it — that the contract being entire there can be no apportionment —and that there being an express contract no other can be implied, even upon the subsequent performance of service — is not properly applicable to this species of contract, where a beneficial service has been actually performed ; for we have abundant reason to believe, that the general understanding of the community is, that the the hired laborer shall be entitled to compensation for the service actually performed, though he do not continue the entire term contracted for, and such contracts must be presumed to be made with reference to that understanding, unless an express stipulation shows the contrary.

Where a beneficial service has been performed and received, therefore, under contracts of this kind, the mutual agreements cannot be considered as going to the whole of the consideration, so as to make them mutual conditions, the one precedent to the other, without a specific proviso to that effect. 1 H. Black. 213, note, Boone v Eyre; 6 D. & E. 570, Campbell v. Jones; 10 East, 295, Ritchie v. Atkinson; 4 Taunt. 745, Burn v. Miller.

It is easy, if parties so choose, to provide by an ex-prés agreement that nothing shall be earned, if the laborer leaves his employer without having performed the whole service contemplated, and then there can be in *494pretence for a recovery if he voluntarily deserts the service before the expiration of the time.

The amount, however, for which the employer ought to be charged, where the laborer abandons Ins contract, is only the reasonable worth, or the amount of advantage lie receives upon the whole transaction, (ante 15, Wad-leigh v. Sutton,) and, in estimating the value of the labor, the contract price for the service cannot be exceeded. 7 Green. 78; 4 Wendell, 285, Dubois v. Delaware & Hudson Canal Company; 7 Wend. 121, Koon v. Greenman.

If a person makes a contract fairly he is entitled to have it fully performed, and if this is not done he is entitled to damages. He ¡nay maintain a suit to recover the amount of damage sustained by the non performance.

The benefit and advantage which the party takes by the labor, therefore, is the amount of value which lie receives, if any, after deducting the amount of damage ; and if he elects to put this in defence he is entitled so to do, and the implied promise which the law will raise, in such case, is to pay such amount of the stipulated price for the whole labor, as remains after deducting what it would cost to procure a completion of the residue of the service, and also any damage which has been sustained by reason of the non fulfilment of the contract.

If in such case it be found that the damages are equal to, or greater than the amount of the labor performed, so that the employer, having a right to the full performance of the contract, has not upon the whole case received a beneficial service, the plaintiff cannot recover.

This rule, by binding the employer to pay the value of the service he actually receives, and the laborer to answer in damages where he does not complete the entire contract, will leave no temptation to the former to drive the laborer from his service, near the close of his term, by ill treatment, in order to escape from payment ; nor to the latter to desert his service before the stipulated time, without a sufficient reason ; and it will in most in*495stances settle the whole controversy in one action, and prevent a multiplicity of suits and cross actions.

There may be instances, however, where the damage occasioned is much greater than the value of the labor performed, and if the party elects to permit himself to be charged for the value of the labor, without interposing the damages in defence, he is entitled to do so, and may have an action to recover his damages for the non-perlormanee, whatever, they may be. 1 Mason’s Rep. Crowninshield v. Robinson.

And he nuiy commence such action at any time after the contract is broken, notwithstanding no suit has been instituted against him ; but if he elects to have the damages considered in the action against him, he must be understood as conceding that they are not to be extended beyond the amount of what he has received, and he cannot afterwards sustain an action for farther damages.

Applying the principles thus laid down, to this case, the plaintiff is entitled to judgment on the verdict.

The defendant sets up a mere breach of the contract in defence of the action, but this cannot avail him. He does not appear to have offered evidence to show that he was damnified by such breach, or to have asked that a deduction should be made upon that account. The direction to the jury was therefore correct, that the plaintiff was entitled to recover as much as the labor performed was reasonably worth, and the jury appear to have allowed a pro rata compensation, for the time which the plaintiff labored in the defendant’s service.

As the defendant has not claimed or had any adjustment of damages, for the breach of the contract, in this action, if he has actually sustained damage he is still entitled to a suit to recover the amount.

Whether it is not necessary, in cases of this kind, that notice should be given to the employer that the contract is abandoned, with an offer of adjustment and demand of payment ; and whether the laborer must not wait until *496the time when the money would have been clue according- to the contract, before commencing an action, (5 B. & P. 61) are questions not necessary to be settled in this case, no objections of that nature having been taken here,

Judgment on the verdict.

OPINION OF THE COURT

Per Curiam.

The issue in this appeal is whether a plaintiff, in an action *260for breach of contract, may recover loss of prospective profits for its contemplated 20-year operation of a domed stadium which was to be constructed by defendant County of Erie (County).

On August 8, 1969, pursuant to a duly adopted resolution of its legislature, the County of Erie entered into a contract with Kenford Company, Inc. (Kenford) and Dome Stadium, Inc. (DSI) for the construction and operation of a domed stadium facility near the City of Buffalo. The contract provided that construction of the facility by the County would commence within 12 months of the contract date and that a mutually acceptable 40-year lease between the County and DSI for the operation of said facility would be negotiated by the parties and agreed upon within three months of the receipt by the County of preliminary plans, drawings and cost estimates. It was further provided that in the event a mutually acceptable lease could not be agreed upon within the three-month period, a separate management contract between the County and DSI, as appended to the basic agreement, would be executed by the parties, providing for the operation of the stadium facility by DSI for a period of 20 years from the completion of the stadium and its availability for use.

Although strenuous and extensive negotiations followed, the parties never agreed upon the terms of a lease, nor did construction of a domed facility begin within the one-year period or at any time thereafter. A breach of the contract thus occurred and this action was commenced in June 1971 by Kenford and DSI.

Prolonged and extensive pretrial and preliminary proceedings transpired throughout the next 10 years, culminating with the entry of an order which affirmed the grant of summary judgment against the County on the issue of liability and directed a trial limited to the issue of damages (Kenford Co. v County of Erie, 88 AD2d 758, lv dismissed 58 NY2d 689). The ensuing trial ended some nine months later with a multimillion dollar jury verdict in plaintiffs’ favor. An appeal to the Appellate Division resulted in a modification of the judgment. That court reversed portions of the judgment awarding damages for loss of profits and for certain out-of-pocket expenses incurred, and directed a new trial upon other issues (Kenford Co. v County of Erie, 108 AD2d 132). On appeal to this court, we are concerned only with that portion *261of the verdict which awarded DSI money damages for loss of prospective profits during the 20-year period of the proposed management contract, as appended to the basic contract. That portion of the verdict was set aside by the Appellate Division and the cause of action dismissed. The court concluded that the use of expert opinion to present statistical projections of future business operations involved the use of too many variables to provide a rational basis upon which lost profits could be calculated and, therefore, such projections were insufficient as a matter of law to support an award of lost profits. We agree with this ultimate conclusion, but upon different grounds.

Loss of future profits as damages for breach of contract have been permitted in New York under long-established and precise rules of law. First, it must be demonstrated with certainty that such damages have been caused by the breach and, second, the alleged loss must be capable of proof with reasonable certainty. In other words, the damages may not be merely speculative, possible or imaginary, but must be reasonably certain and directly traceable to the breach, not remote or the result of other intervening causes (Wakeman v Wheeler & Wilson Mfg. Co., 101 NY 205). In addition, there must be a showing that the particular damages were fairly within the contemplation of the parties to the contract at the time it was made (Witherbee v Meyer, 155 NY 446). If it is a new business seeking to recover for loss of future profits, a stricter standard is imposed for the obvious reason that there does not exist a reasonable basis of experience upon which to estimate lost profits with the requisite degree of reasonable certainty (Cramer v Grand Rapids Show Case Co., 223 NY 63; 25 CJS, Damages, § 42 [b]).

These rules must be applied to the proof presented by DSI in this case. We note the procedure for computing damages selected by DSI was in accord with contemporary economic theory and was presented through the testimony of recognized experts. Such a procedure has been accepted in this State and many other jurisdictions (see, De Long v County of Erie, 60 NY2d 296). DSI’s economic analysis employed historical data, obtained from the operation of other domed stadiums and related facilities throughout the country, which was then applied to the results of a comprehensive study of the marketing prospects for the proposed facility in the Buffalo area. The quantity of proof is massive and, unquestionably, represents business and industry’s most advanced and sophisticated *262method for predicting the probable results of contemplated projects. Indeed, it is difficult to conclude what additional relevant proof could have been submitted by DSI in support of its attempt to establish, with reasonable certainty, loss of prospective profits. Nevertheless, DSI’s proof is insufficient to meet the required standard.

The reason for this conclusion is twofold. Initially, the proof does not satisfy the requirement that liability for loss of profits over a 20-year period was in the contemplation of the parties at the time of the execution of the basic contract or at the time of its breach (see, Chapman v Fargo, 223 NY 32; 36 NY Jur 2d, Damages, §§39, 40, at 66-70). Indeed, the provisions in the contract providing remedy for a default do not suggest or provide for such a heavy responsibility on the part of the County. In the absence of any provision for such an eventuality, the commonsense rule to apply is to consider what the parties would have concluded had they considered the subject. The evidence here fails to demonstrate that liability for loss of profits over the length of the contract would have been in the contemplation of the parties at the relevant times.

Next, we note that despite the massive quantity of expert proof submitted by DSI, the ultimate conclusions are still projections, and as employed in the present day commercial world, subject to adjustment and modification. We of course recognize that any projection cannot be absolute, nor is there any such requirement, but it is axiomatic that the degree of certainty is dependent upon known or unknown factors which form the basis of the ultimate conclusion. Here, the foundations upon which the economic model was created undermine the certainty of the projections. DSI assumed that the facility was completed, available for use and successfully operated by it for 20 years, providing professional sporting events and other forms of entertainment, as well as hosting meetings, conventions and related commercial gatherings. At the time of the breach, there was only one other facility in this country to use as a basis of comparison, the Astrodome in Houston. Quite simply, the multitude of assumptions required to establish projections of profitability over the life of this contract require speculation and conjecture, making it beyond the capability of even the most sophisticated procedures to satisfy the legal requirements of proof with reasonable certainty.

The economic facts of life, the whim of the general public *263and the fickle nature of popular support for professional athletic endeavors must be given great weight in attempting to ascertain damages 20 years in the future. New York has long recognized the inherent uncertainties of predicting profits in the entertainment field in general (see, Broadway Photoplay Co. v World Film Corp., 225 NY 104) and, in this case, we are dealing, in large part, with a new facility furnishing entertainment for the public. It is our view that the record in this case demonstrates the efficacy of the principles set forth by this court in Cramer v Grand Rapids Show Case Co. (223 NY 63, supra), principles to which we continue to adhere. In so doing, we specifically reject the "rational basis” test enunciated in Perma Research & Dev. Co. v Singer Co. (542 F2d 111, cert denied 429 US 987) and adopted by the Appellate Division.

Accordingly, that portion of the order of the Appellate Division being appealed from should be affirmed.

Chief Judge Wachtler and Judges Meyer, Alexander, Titone and Kane* concur in Per Curiam opinion; Judges Simons, Kaye and Hancock, Jr., taking no part.

Order insofar as appealed from affirmed, with costs.

POSNER, Circuit Judge.

These appeals bring before us various facets of a dispute between Northern Indiana Public Service Company (NIPSCO), an electric utility in Indiana, and Carbon County Coal Company, a partnership that until recently owned and operated a coal mine in Wyoming. In 1978 NIPSCO and Carbon County signed a contract whereby Carbon County agreed to sell and NIPSCO to buy approximately 1.5 million tons of coal every year for 20 years, at a price of $24 a ton subject to various provisions for escalation which by 1985 had driven the price up to $44 a ton.

NIPSCO’s rates are regulated by the Indiana Public Service Commission. In 1983 NIPSCO requested permission to raise its rates to reflect increased fuel charges. Some customers of NIPSCO opposed the increase on the ground that NIP-SCO could reduce its overall costs by buying more electrical power from neighboring utilities for resale to its customers and producing less of its own power. Although the Commission granted the requested increase, it directed NIPSCO, in orders issued in December 1983 and February 1984 (the “economy purchase orders”), to make a good faith effort to find, and wherever possible buy from, utilities that would sell electricity to it at prices lower than its costs of internal generation. The Commission added ominously that “the adverse effects of entering into long-term coal supply contracts which do not allow for renegotiation and are not requirement contracts, is a burden which must rest squarely on the shoulders of NIPSCO management.” Actually the contract with Carbon County did provide for renegotiation of the contract price — but one-way renegotiation in favor of Carbon County; the price fixed in the contract (as adjusted from time to time in accordance with the escalator provisions) was a floor. And the contract was indeed not a requirements contract: it specified the exact amount of coal that NIPSCO must take over the 20 years during which the contract was to remain in effect. NIP-SCO was eager to have an assured supply of low-sulphur coal and was therefore willing to guarantee both price and quantity.

Unfortunately for NIPSCO, as things turned out it was indeed able to buy electricity at prices below the costs of generating electricity from coal bought under the contract with Carbon County; and because of the “economy purchase orders,” of which it had not sought judicial review, NIPSCO could not expect to be allowed by the Public Service Commission to recover in its electrical rates the costs of buying coal from Carbon County. NIPSCO therefore decided to stop accepting coal deliveries from Carbon County, at least for the time being; and on April 24, 1985, it brought this diversity suit against Carbon County in *268a federal district court in Indiana, seeking a declaration that it was excused from its obligations under the contract either permanently or at least until the economy purchase orders ceased preventing it from passing on the costs of the contract to its ratepayers. In support of this position it argued that the contract violated section 2(c) of the Mineral Lands Leasing Act of 1920, 30 U.S.C. § 202, because of Carbon County’s affiliation with a railroad (Union Pacific), and that in any event NIPSCO’s performance was excused or suspended— either under the contract’s force majeure clause or under the doctrines of frustration or impossibility — by reason of the economy purchase orders.

On May 17, 1985, Carbon County counterclaimed for breach of contract and moved for a preliminary injunction requiring NIPSCO to continue taking delivery under the contract. On June 19, 1985, the district judge granted the preliminary injunction, from which NIPSCO has appealed. Also on June 19, rejecting NIPSCO’s argument that it needed more time for pretrial discovery and other trial preparations, the judge scheduled the trial to begin on August 26, 1985. Trial did begin then, lasted for six weeks, and resulted in a jury verdict for Carbon County of $181 million. The judge entered judgment in accordance with the verdict, rejecting Carbon County’s argument that in lieu of damages it should get an order of specific performance requiring NIPSCO to comply with the contract. Upon entering the final judgment the district judge dissolved the preliminary injunction, and shortly afterward the mine — whose only customer was NIP-SCO — shut down. NIPSCO has appealed from the damage judgment, and Carbon County from the denial of specific performance and from the district judge’s order staying execution of the damage judgment without requiring NIPSCO to post a bond guaranteeing payment of the judgment should NIPSCO lose on appeal.

NIPSCO’s appeal from the grant of the preliminary injunction is moot, the injunction having been dissolved last October when the final judgment was entered. Lifting a preliminary injunction does not always make an appeal from the grant of the injunction moot; if the injunction was improper, the defendant may be entitled to damages. See Fed.R.Civ.P. 65(c); Coyne-Delany Co. v. Capital Development Bd., 717 F.2d 385 (7th Cir.1983). But all that NIPSCO is asking for in appealing from the grant of the preliminary injunction is that the injunction be dissolved, and that request is moot. Later we shall see that another issue in the case, relating to the judge’s instructions to the jury on force majeure, is also moot.

We are left with the following issues to decide: (1) whether the district judge abused his discretion in refusing to give NIPSCO more time to prepare for trial, (2) whether the contract was unenforceable as a violation of the Mineral Lands Leasing Act, (3) whether NIPSCO’s obligations under the contract were excused or suspended by virtue of either the force majeure clause or (4) the doctrines of frustration or impracticability, (5) whether Carbon County was entitled to specific performance of the contract, and (6) whether NIPSCO should be required to post a bond in order to be allowed to stave off the execution of the damage judgment until the appellate process is over.

1. When he issued the preliminary injunction, two months into the case, the district judge scheduled the trial to begin in two months. This was a tight deadline for the completion of pretrial discovery, though NIPSCO is wrong to argue that it violates a local rule of the Northern District of Indiana. Rule 12(d) provides that “in all civil cases except in patent, antitrust, and trade-mark cases, all discovery shall be completed within five months after the case is at issue.” Five months is the maximum, not the minimum. With somewhat greater force NIPSCO argues that this case is extraordinary, even though it is not a patent, antitrust, or trademark case. Although the issues are no more complex than they would be if the contract had been for $1 million rather than $1 billion (the *269estimated amount that NIPSCO would have owed over the life of the contract if the contract had not been cancelled), the large stakes justified a more careful and thorough preparation than if the stakes had been smaller: the consequences of an erroneous judgment were greater, so the optimal expenditure on avoiding error was greater. NIPSCO particularly complains about its inability to conduct thorough discovery of Carbon County’s relationship to the Union Pacific Railroad — a relationship that as we shall see is the basis of NIP-SCO’s defense based on the Mineral Lands Leasing Act — and of Carbon County’s theory of damages. And the only reason the district judge gave for drastically compressing the pretrial period was that his criminal trial calendar was so crowded (with cases not deferrable, because of the requirements of the Speedy Trial Act) that if he did not try the case in August 1985 he would have to put it over to April 1986—which still would have been only a year after the complaint was filed.

Nevertheless we do not think the district judge abused his discretion in refusing to postpone the trial. Matters of trial management are for the district judge; we intervene only when it is apparent that the judge has acted unreasonably. The occasions for intervention are rare. At a time when a combination of very heavy caseloads with the pressure exerted by the Speedy Trial Act to take criminal cases out of order and try them first makes managing federal district courts’ dockets trickier than ever before, district judges must be allowed considerable leeway in scheduling civil cases, and therefore in denying continuances that would disrupt their schedules. Afram Export Corp. v. Metallurgiki Halyps, S.A., 772 F.2d 1358, 1366 (7th Cir.1985); see also Fontenot v. Upjohn Co., 780 F.2d 1190, 1193 (5th Cir.1986); cf. Kagan v. Caterpillar Tractor Co., 795 F.2d 601, 608-09 (7th Cir.1986).

Of course some cases are so immense that it would be unrealistic to insist on trial after only two months for pretrial discovery, or even to insist on rigid adherence to deadlines in local rules. But this is not such a case. To begin with, the two-month figure is misleading. NIPSCO had decided in 1984 to stop accepting coal deliveries from Carbon County and had retained late in that year — nine months or more before the trial began — the large Chicago law firm that would later conduct the trial with the help of the much smaller Indiana firm that is NIPSCO’s regular counsel. The Chicago firm has between 40 and 50 lawyers in its trial department. The timing of the suit was in the control of NIPSCO and its lawyers, and they could do much of their trial preparation before filing suit. True, they could not have conducted discovery then. Discovery is sometimes allowed before suit is filed, see Fed.R.Civ.P. 27(a), but not in the circumstances of this case. The defendants could, however, have begun discovery when they filed their suit, four months before the beginning of the trial and almost six months before the end; the latter is the relevant interval, because the judge interrupted the trial to allow NIP-SCO to complete its discovery on the damage issue.

No case holds that the denial of a continuance in such circumstances is reversible error. The common element in cases that reverse such denials is the existence of changed circumstances to which a party cannot reasonably be expected to adjust without an extension of time. In Sutherland Paper Co. v. Grant Paper Box Co., 183 F.2d 926, 931 (3d Cir.1950), the denial of certain pretrial motions just two weeks before the trial of a complex patent case, combined with the illness of the inventor, made it impossible for the defendants to complete their pretrial preparation without a continuance that the judge denied. In Smith-Weik Machinery Corp. v. Murdock Machine & Engineering Co., 423 F.2d 842 (5th Cir.1970), the defendant’s principal counsel became ill on the eve of trial, and its local counsel was not sufficiently well informed about the facts and pertinent law to conduct the trial by himself on such short notice. In Fenner v. Dependable Trucking Co., 716 F.2d 598, 600-02 (9th Cir.1983), the judge refused to grant a con*270tinuanee to enable the defendants’ expert to rebut evidence newly discovered by the plaintiff. Other “surprise” cases are Wells v. Rushing, 755 F.2d 376, 380-81 (5th Cir.1985); Menendez v. Perishable Distributors, Inc., 763 F.2d 1374, 1379-80 (11th Cir.1985), and Shelak v. White Motor Co., 581 F.2d 1155, 1159 (5th Cir.1978) (and see Judge Rubin’s powerful dissent, id. at 1161-64). The element of changed circumstances is missing here.

Lack of precedent need not defeat NIPSCO’s argument for reversal; NIPSCO can appeal to first principles, and ask us to forge new ground. But this we shall not do, when NIPSCO is unable to show how it was prejudiced by the compressed period for discovery and pretrial preparation, given the amount of time it had to plan the case before it filed suit and the large staff of lawyers at its disposal to conduct discovery once the suit began. On the view that we take of its defense of illegality, nothing NIPSCO could reasonably have been expected to obtain from additional discovery on the linkage between Carbon County and the Union Pacific Railroad could change the result of the case. As for damages, it is noteworthy both that NIP-SCO does not complain in this court about the size of the jury verdict and that in April 1984 it had estimated that cancelling the contract would cost it $300 million in damages if Carbon County could show that the cancellation was a breach of contract. At argument NIPSCO’s counsel acknowledged that the compression of the time for pretrial preparation was as harmful to Carbon County as to itself. Even if as a matter of abstract justice the judge should have given NIPSCO more time for conducting discovery, his refusal to do so would not be reversible error unless there was a reasonable probability that it caused NIPSCO to lose the case, see Fed.R.Civ.P. 61; Fontenot v. Upjohn Co., supra, 780 F.2d at 1194; Menendez v. Perishable Distributors, Inc., supra, 763 F.2d at 1380, and there is no indication of that. Finally, in arguing against the grant of a preliminary injunction NIPSCO had said that having to continue taking coal under the contract would be very costly to it; and if so it should have welcomed an early trial. Business Ass’n of University City v. Landrieu, 660 F.2d 867, 877-78 (3d Cir.1981), upheld the denial of a continuance in parallel circumstances.

Justified public concern with the expenses and delays of modern litigation requires reexamination of traditional attitudes about the big case — one of which is that such a case cannot be tried without leisurely pretrial preparation. Despite the abbreviated period for pretrial discovery in the present case, both parties had an adequate opportunity, which they used, to obtain all the evidence they needed for the trial and to scrutinize the opponent’s evidence closely. Rather than believing that the district judge abused the broad discretion that our system gives trial judges in the management of litigation, we find nothing to criticize in the dispatch with which this big case was brought to judgment.

2. Section 2(c) of the Mineral Lands Leasing Act of 1920 provides in pertinent part that “no company or corporation operating a common-carrier railroad shall be given or hold a permit or lease under the provisions of this chapter [relating to federal lands] for any coal deposits except for its own use for railroad purposes____” Oddly in this litigious age, no reported decision has interpreted section 2(c) in the 66 years since its enactment. NIPSCO contends that if the statute is not to be made a dead letter, it must be read to forbid a railroad’s affiliate to hold a mineral lease or permit on federal lands. Roughly 15 percent of Carbon County’s projected output for the contract with NIPSCO was to come from federal lands that Carbon County had a permit to mine, and Carbon County is a partnership of two firms (each with a half interest in the partnership), Dravo Coal Company and Rocky Mountain Energy Company, the latter a wholly owned subsidiary of the Union Pacific Corporation, whose principal subsidiary is the Union Pacific Railroad Company.

*271When the contract was made back in 1978, the Department of Interior took the view that section 2(c) did not require the automatic piercing of corporate veils, and hence did not invalidate leases by subsidiaries or other affiliates of railroads unless the affiliate was an “alter ego” of the railroad, meaning that their corporate separateness was a paper formality with no business or economic significance. In 1980 the Department of Justice, while recommending to Congress that section 2(c) be repealed as an anachronism, advised it that the section does reach affiliates, and in 1982 the Department of the Interior adopted this interpretation, though only for prospective application. Railroad Affiliates and Coal Leasing, 89 I.D. 610 (1982).

No one doubts that section 2(c) reaches a lease by a railroad’s alter ego. The statutory language, “company ... operating a ... railroad,” can without contortion be interpreted to cover a situation where the railroad sets up a dummy corporation to hold a mineral lease on federal lands and places control of the mining operations in the railroad’s management; if such facile evasions could not be prevented, the statute would have very little consequence. By way of comparison consider the effort of the Federal Communications Commission to make its requirement that communications common carriers conduct their terminal equipment business through separate subsidiaries a meaningful one by forbidding the subsidiary and the carrier to share staff or facilities. See Illinois Bell Tel Co. v. FCC, 740 F.2d 465, 473-74 (7th Cir.1984). But if section 2(c) merely covers the railroad plus its alter egos, NIPSCO must lose, for Carbon County is not the alter ego of the Union Pacific Railroad. In recent years the railroad has generated only about half of Union Pacific Corporation’s total revenues; and the corporation’s Rocky Mountain subsidiary is only a 50 percent partner in Carbon County and the other partner both is unrelated to the Union Pacific family and was an active partner in the venture. Even if we treated the holding company as the alter ego of the railroad, we could not treat Carbon County so, if only because an equal partner cannot call the tune when the other partner is active, knowledgeable, and independent.

But does section 2(c) go beyond alter egos to embrace affiliates of railroads? It was enacted out of a fear that if railroads were allowed to own coal mines they would discriminate in transportation against competing coal mines which depended on rail transportation; this was thought to have happened in the anthracite mine fields of the eastern United States. See 58 Cong. Rec. 4739 (1919) (remarks of Sen. LaFol-lette); United States v. Reading Co., 253 U.S. 26, 40 S.Ct. 425, 64 L.Ed. 760 (1920). Since the railroads owned their coal mines through subsidiaries rather than directly, it is indeed true that the purpose of section 2(c) could not be fully achieved unless “company ... operating a ... railroad” were interpreted to mean “company ... affiliated with a ... railroad.” And strained as this interpretation might seem from a linguistic standpoint, other strained statutory interpretations have been adopted where necessary to prevent a statute’s evident, but imperfectly expressed, purpose, whether of inclusion or exclusion, from being too easily defeated. Yet so patent is the oversight — given the corporate form in which the eastern railroads had held their coal mines, the narrow interpretation the Supreme Court had given the parallel “commodities clause” of the Interstate Commerce Act in United States ex rel. Attorney Gen’l v. Delaware & Hudson Co., 213 U.S. 366, 413-14, 29 S.Ct. 527, 538, 53 L.Ed. 836 (1909), and the fact that an earlier version of the bill that became section 2(c) referred explicitly to subsidiaries and affiliates — that one suspects it was deliberate. The proponents of cracking down on railroads’ ownership of coal mines must not have had the votes to get a fully effective statute. We note that United States v. Reading Co., supra, 253 U.S. at 60-62, 40 S.Ct. at 433, as interpreted in United States v. Elgin, Joliet & Eastern Ry., 298 U.S. 492, 501-02, 56 S.Ct. 841, 844, 80 L.Ed. 1300 (1936), became the source of an “alter ego” interpretation of the com*272modities clause which the Department of the Interior then borrowed for section 2(c) of the Mineral Lands Leasing Act. Perhaps no broader interpretation of the section can be squared with its language and history.

Support for this conclusion is found in the peculiar pattern of railroad land holdings in the West, described in Leo Sheep Co. v. United States, 440 U.S. 668, 672-77, 99 S.Ct. 1403, 1406-08, 59.L.Ed.2d 677 (1979). Beginning in 1862, Congress, in an effort to induce the Union Pacific to build a transcontinental railroad, had granted the Union Pacific large tracts of land along the projected railroad right of way. To reduce the cost to the government, Congress configured the grant in a checkerboard pattern: the Union Pacific got alternate blocks of land, and as a result every block granted to the Union Pacific was surrounded by federal land and every block of federal land was surrounded by Union Pacific land. The idea was that the federal government would benefit equally with Union Pacific from any appreciation in land values that was brought about by the creation of the railroad. The checkerboard pattern created problems for coordinated development, and Congress in 1920 may not have wanted to inhibit the development of contiguous lands by preventing the Union Pacific from obtaining mineral leases for subsidiaries already engaged in mining on the railroad’s own land. The complications caused by the checkerboard grant, as well as the obsolescence of the monopoly concerns (possibly exaggerated even in 1920) that had led to the passage of section 2(c), are what persuaded the Justice Department to recommend repealing the statute.

All this leaves in doubt whether section 2(c) reaches affiliate relationships short of alter egoism (on the distinction between a mere affiliate and an alter ego in other legal settings see, e.g., Crest Tankers, Inc. v. National Maritime Union, 796 F.2d 234 (8th Cir.1986)); but even if it does, the relationship here is so attenuated that one may doubt whether Carbon County should even be called an affiliate of the Union Pacific Railroad. And even if Carbon County should not have been granted a permit to mine coal on federal land, this would not allow NIPSCO to avoid its obligations under the contract. To begin with, assuming there is a violation of section 2(c) lurking somewhere in the background, the contract itself is not illegal. The statute does not regulate sales of coal mined on federal land. It prohibits railroads from holding leases or permits to mine coal on those lands (other than for the railroad’s own use — an anachronistic exception in the age of the diesel), but says nothing about sales of the coal. And it is not the sale that conceivably offends the statute but the seller’s affiliation — a problem that could be solved, without termination of the contract, by Union Pacific Corporation’s selling Rocky Mountain Energy Company. Moreover, Carbon County’s mine is mainly on private land. Apparently there is not enough coal on that land to supply the entire contract; and the contract requires Carbon County to supply the coal for the contract from this mine. Nevertheless the contract could not be thought illegal in its entirety. At most NIPSCO could claim an abatement of some of the contract damages, representing profits that would have accrued from the sale of the coal mined on public land. NIPSCO claimed no such abatement.

Since this is not a case where the contract itself is illegal — as it would be if it were a contract in restraint of trade and therefore a violation of section 1 of the Sherman Act, or a contract to commit a bank robbery and therefore a criminal conspiracy — it is not governed by Kaiser Steel Corp. v. Mullins, 455 U.S. 72, 77-83, 102 S.Ct. 851, 856-59, 70 L.Ed.2d 833 (1982). Kaiser agreed to make extra contributions to a union welfare fund if Kaiser failed to adhere to an allegedly illegal boycott; the Court sustained the defense of illegality to a suit to enforce the agreement. The extra contributions were in effect a penalty for abandoning the boycott; the underlying agreement that the penalty was designed to enforce was thus an agreement to participate in the boycott. See also Somerset *273 Importers, Ltd. v. Continental Vintners, 790 F.2d 775 (9th Cir.1986). In contrast, the contract in this case is not “intrinsically illegal,” Trustees of the Operative Plasterers’ and Cement Masons’ Local Union Officers & Employees Pension Fund v. Journeyman Plasterers’ Protective & Benevolent Soc’y, Local Union No. 5, 794 F.2d 1217, 1220 (7th Cir.1986); and the defense of illegality does not come into play just because a party to a lawful contract (here a contract to supply coal to an electric utility) commits unlawful acts to carry out his part of the bargain. See, e.g., Michigan Millers Mutual Fire Ins. Co. v. Canadian Northern Ry., 152 F.2d 292, 297 (8th Cir.1945); Drost v. Professional Building Service Corp., 153 Ind.App. 273, 279, 286 N.E.2d 846, 850 (1972).

Second, supposing that the contract does violate section 2(c) of the Mineral Lands Leasing Act, this does not necessarily make it unenforceable. This issue, too, is one of federal rather than state law, though we have no reason to think Indiana law would require a different resolution of it; federal and state law on the contract defense of illegality — the latter well described in Farnsworth, Contracts §§ 5.5, 5.6 (1982) — seem quite similar. When the statute is federal, federal law determines not only whether the statute was violated but also, if so, and assuming the statute itself is silent on the matter, the effect of the violation on the enforceability of the contract. Walsh v. Schlecht, 429 U.S. 401, 407-08, 97 S.Ct. 679, 684-85, 50 L.Ed.2d 641 (1977); Kelly v. Kosuga, 358 U.S. 516, 519, 79 S.Ct. 429, 431, 3 L.Ed.2d 475 (1959); Sola Electric Co. v. Jefferson Electric Co., 317 U.S. 173, 176, 63 S.Ct. 172, 174, 87 L.Ed. 165 (1942); cf. Kaiser Steel Corp. v. Mullins, supra, 455 U.S. at 77, 102 S.Ct. at 856.

But when we ask what is the federal rule on illegality as a contract defense, we find, alas, that the course of decision has not run completely true. Compare, for example, the hard line taken in government-contract cases such as United States v. Mississippi Valley Generating Co., 364 U.S. 520, 563-66, 81 S.Ct. 294, 316-17, 5 L.Ed.2d 268 (1961), where the defense seems almost automatic (see also Comdisco, Inc. v. United States, 756 F.2d 569, 576 (7th Cir.1985); but cf. United States v. Medico Industries, Inc., 784 F.2d 840, 845 (7th Cir.1986)), with the very soft line taken in antitrust cases such as Kelly v. Kosuga, supra, and Delta Marina, Inc. v. Plaquemine Oil Sales, Inc., 644 F.2d 455, 458-59 (5th Cir.1981); but cf. Kaiser Steel Corp. v. Mullins, supra, 455 U.S. at 79-82, 102 S.Ct. at 857-59. The best generalization possible is that the defense of illegality, being in character if not origins an equitable and remedial doctrine, is not automatic but requires (as NIPSCO’s counsel acknowledged at argument) a comparison of the pros and cons of enforcement. See Jackson Purchase Rural Electric Coop. Ass’n v. Local Union 816, Int’l Brotherhood of Electrical Workers, 646 F.2d 264, 267 (6th Cir.1981); cf. 15 Williston, A Treatise on the Law of Contracts § 1767, at pp. 264-65 (Jaeger 3d ed. 1972).

There are, after all, statutory remedies (e.g., cancellation of the lease or permit, see 30 U.S.C. § 188(b)) for violations of section 2(c); the question is whether there should be an additional, a judge-made, remedy. To decide whether there should be, we must consider the reciprocal dangers of overdeterrenee and underdeterrence. Applied too strictly, the doctrine that makes the unenforceability of a contract an additional remedy for the violation of a statute can produce a disproportionately severe sanction; and the overdeterrence of illegality is as great a danger to freedom and prosperity as underdeterrence. The benefits of enforcing the tainted contract — benefits that lie in creating stability in contract relations and preserving reasonable expectations — must be compared with the costs in forgoing the additional deterrence of behavior forbidden by statute that is brought about by refusing to let the violator enforce the contract.

The balance in this case favors enforcement. This makes it irrelevant whether the district judge improperly instructed the *274jury that in order to uphold the defense of illegality, it must find both that Carbon County was an alter ego of the railroad and that NIPSCO had been injured by the violation of section 2(c), or even whether the contract itself could be viewed as illegal under any interpretation of the statute. The interpretation of section 2(c) by the Departments of Justice and the Interior as embracing mere affiliates reversed the Department of the Interior’s previous interpretation and was sufficiently unexpected to lead the Department to make the new interpretation prospective only — and the contract in this case had been signed years earlier. In any event the violation of section 2(c) (if any) was trivial given the attenuated linkage between the Union Pacific Railroad and Carbon County, and so far as appears completely harmless. No competitor of Union Pacific in the railroad business, and no competitor of Rocky Mountain Energy Company in the coal business who might be dependent on Union Pacific to transport his coal — no competitor of any member of the Union Pacific “family,” however broadly defined — has complained that Carbon County is violating section 2(c). Nor has the Department of the Interior or the Department of Justice. Nor has any customer of any of the entities involved (however peripherally) in this lawsuit. Compare National Licorice Co. v. NLRB, 309 U.S. 350, 364-66, 60 S.Ct. 569, 577-78, 84 L.Ed. 799 (1940); Republic Airlines, Inc. v. United Air Lines, Inc., 796 F.2d 526, 528 (D.C.Cir.1986). Only NIPSCO complains.

Section 2(c) is an anachronism — a regulatory statute on which the sun set long ago. It could serve as Exhibit A to Dean Cala-bresi’s proposal that courts be empowered to invalidate obsolete statutes without having to declare them unconstitutional. See Calabresi, A Common Law for the Age of Statutes (1982). We do not believe that we have the power to declare a constitutional statute invalid merely because we, or for that matter everybody, think the statute has become obsolete. But the question in this case is not whether section 2(c) is enforceable but whether an alleged violation of the statute makes a contract unenforceable, and the obsolescence of the statute may be relevant to that determination.

NIPSCO does not argue that Carbon County’s alleged violation of the statute hurt NIPSCO or that invalidating this contract under section 2(c) would help anyone, anywhere, at any time. The only consequence, other than to the parties to this suit, would be to throw a cloud of uncertainty over hundreds, perhaps thousands, of contracts for the supply of coal by firms affiliated with railroads, and to inject uncertainty into the contracting process generally. Persons negotiating contracts would have to worry about whether their contract might someday be found to have violated some old, little-known, and newly reinterpreted statute. Lawyers would benefit from the need to do more legal research before signing a contract, but no one else would. We conclude that the Mineral Lands Leasing Act is not a defense to the enforcement of this contract.

3. The contract permits NIPSCO to stop taking delivery of coal “for any cause beyond [its] reasonable control ... including but not limited to ... orders or acts of civil ... authority ... which wholly or partly prevent ... the utilizing ... of the coal.” This is what is known as a force majeure clause. See, e.g., Northern Illinois Gas Co. v. Energy Coop., Inc., 122 Ill.App.3d 940, 949-52, 78 Ill.Dec. 215, 223-24, 461 N.E.2d 1049, 1057-58 (1984). NIP-SCO argues that the Indiana Public Service Commission’s “economy purchase orders” prevented it, in whole or part, from using the coal that it had agreed to buy, and it complains that the district judge instructed the jury incorrectly on the meaning and application of the clause. The complaint about the instructions is immaterial. The judge should not have put the issue of force majeure to the jury. It is evident that the clause was not triggered by the orders.

All that those orders do is tell NIPSCO it will not be allowed to pass on fuel costs to its ratepayers in the form of higher rates if *275it can buy electricity cheaper than it can generate electricity internally using Carbon County’s coal. Such an order does not “prevent,” whether wholly or in part, NIP-SCO from using the coal; it just prevents NIPSCO from shifting the burden of its improvidence or bad luck in having incorrectly forecasted its fuel needs to the backs of the hapless ratepayers. The purpose of public utility regulation is to provide a substitute for competition in markets (such as the market for electricity) that are naturally monopolistic. Suppose the market for electricity were fully competitive, and unregulated. Then if NIPSCO signed a long-term fixed-price fixed-quantity contract to buy coal, and during the life of the contract competing electrical companies were able to produce and sell electricity at prices below the cost to NIPSCO of producing electricity from that coal, NIPSCO would have to swallow the excess cost of the coal. It could not raise its electricity prices in order to pass on the excess cost to its consumers, because if it did they would buy electricity at lower prices from NIP-SCO’s competitors. By signing the kind of contract it did, NIPSCO gambled that fuel costs would rise rather than fall over the life of the contract; for if they rose, the contract price would give it an advantage over its (hypothetical) competitors who would have to buy fuel at the current market price. If such a gamble fails, the result is not force majeure.

This is all the clearer when we consider that the contract price was actually fixed just on the downside; it put a floor under the price NIPSCO had to pay, but the escalator provisions allowed the actual contract prices to rise above the floor, and they did. This underscores the gamble NIPSCO took in signing the contract. It committed itself to paying a price at or above a fixed minimum and to taking a fixed quantity at that price. It was willing to make this commitment to secure an assured supply of low-sulphur coal, but the risk it took was that the market price of coal or substitute fuels would fall. A force majeure clause is not intended to buffer a party against the normal risks of a contract. The normal risk of a fixed-price contract is that the market price will change. If it rises, the buyer gains at the expense of the seller (except insofar as escalator provisions give the seller some protection); if it falls, as here, the seller gains at the expense of the buyer. The whole purpose of a fixed-price contract is to allocate risk in this way. A force majeure clause interpreted to excuse the buyer from the consequences of the risk he expressly assumed would nullify a central term of the contract.

The Indiana Public Service Commission is a surrogate for the forces of competition, and the economy fuel orders are a device for simulating the effects in a competitive market of a drop in input prices. The orders say to NIPSCO, in effect: “With fuel costs dropping, and thus reducing the costs of electricity to utilities not burdened by long-term fixed-price contracts, you had better substitute those utilities’ electricity for your own when their prices are lower than your cost of internal generation. In a freely competitive market consumers would make that substitution; if you do not do so, don’t expect to be allowed to pass on your inflated fuel costs to those consumers.” Admittedly the comparison between competition and regulation is not exact. In an unregulated market, if fuel costs skyrocketed NIPSCO would have a capital gain from its contract (assuming the escalator provisions did not operate to raise the contract price by the full amount of the increase in fuel costs, a matter that would depend on the cause of the increase). This is because its competitors, facing higher fuel costs, would try to raise their prices for electricity, thus enabling NIPSCO to raise its price, or expand its output, or both, and thereby increase its profits. The chance of this “windfall” gain offsets, on an ex ante (before the fact) basis, the chance of a windfall loss if fuel costs drop, though NIPSCO it appears was seeking a secure source of low-sulphur coal rather than a chance for windfall gains. If as is likely the Public Service Commission would require NIPSCO to pass on any capital gain from an advantageous contract to the *276ratepayers (which is another reason for thinking NIPSCO wasn’t after windfall gains — it would not, in all likelihood, have been allowed to keep them), then it ought to allow NIPSCO to pass on to them some of the capital loss from a disadvantageous contract — provided that the contract, when made, was prudent. Maybe it was not; maybe the risk that NIPSCO took was excessive. But all this was a matter between NIPSCO and the Public Service Commission, and NIPSCO did not seek judicial review of the economy purchase orders.

If the Commission had ordered NIPSCO to close a plant because of a safety or pollution hazard, we would have a true case of force majeure. As a regulated firm NIPSCO is subject to more extensive controls than unregulated firms and it therefore wanted and got a broadly worded force majeure clause that would protect it fully (hence the reference to partial effects) against government actions that impeded its using the coal. But as the only thing the Commission did was prevent NIPSCO from using its monopoly position to make consumers bear the risk that NIPSCO assumed when it signed a long-term fixed-price fuel contract, NIPSCO cannot complain of force majeure; the risk that has come to pass was one that NIPSCO voluntarily assumed when it signed the contract.

4. The district judge refused to submit NIPSCO’s defenses of impracticability and frustration to the jury, ruling that Indiana law does not allow a buyer to claim impracticability and does not recognize the defense of frustration. Some background (on which see Farnsworth, Contracts §§ 9.5-9.7 (1982)) may help make these rulings intelligible. In the early common law a contractual undertaking unconditional in terms was not excused merely because something had happened (such as an invasion, the passage of a law, or a natural disaster) that prevented the undertaking. See Paradine v. Jane, Aleyn 26, 82 Eng. Rep. 897 (K.B.1647). Excuses had to be written into the contract; this is the origin of force majeure clauses. Later it came to be recognized that negotiating parties cannot anticipate all the contingencies that may arise in the performance of the contract; a legitimate judicial function in contract cases is to interpolate terms to govern remote contingencies — terms the parties would have agreed on explicitly if they had had the time and foresight to make advance provision for every possible contingency in performance. Later still, it was recognized that physical impossibility was irrelevant, or at least inconclusive; a prom-isor might want his promise to be unconditional, not because he thought he had superhuman powers but because he could insure against the risk of nonperformance better than the promisee, or obtain a substitute performance more easily than the promisee. See Field Container Corp. v. ICC, 712 F.2d 250, 257 (7th Cir.1983); Holmes, The Common Law 300 (1881). Thus the proper question in an “impossibility” case is not whether the promisor could not have performed his undertaking but whether his nonperformance should be excused because the parties, if they had thought about the matter, would have wanted to assign the risk of the contingency that made performance impossible or uneconomical to the promisor or to the promisee; if to the latter, the promisor is excused.

Section 2-615 of the Uniform Commercial Code takes this approach. It provides that “delay in delivery ... by a seller ... is not a breach of his duty under a contract for sale if performance as agreed has been made impracticable by the occurrence of a contingency the non-occurrence of which was a basic assumption on which the contract was made....” Performance on schedule need not be impossible, only infeasible — provided that the event which made it infeasible was not a risk that the promisor had assumed. Notice, however, that the only type of promisor referred to is a seller; there is no suggestion that a buyer’s performance might be excused by reason of impracticability. The reason is largely semantic. Ordinarily all the buyer has to do in order to perform his side of the bargain is pay, and while one can think of all sorts of reasons why, when the time *277came to pay, the buyer might not have the money, rarely would the seller have intended to assume the risk that the buyer might, whether through improvidence or bad luck, be unable to pay for the seller’s goods or services. To deal with the rare case where the buyer or (more broadly) the paying party might have a good excuse based on some unforeseen change in circumstances, a new rubric was thought necessary, different from “impossibility” (the common law term) or “impracticability” (the Code term, picked up in Restatement (Second) of Contracts § 261 (1979)), and it received the name “frustration.” Rarely is it impracticable or impossible for the payor to pay; but if something has happened to make the performance for which he would be paying worthless to him, an excuse for not paying, analogous to impracticability or impossibility, may be proper. See Restatement, supra, § 265, comment a.

The leading case on frustration remains Krell v. Henry, [1903] 2 K.B. 740 (C.A.). Krell rented Henry a suite of rooms for watching the coronation of Edward VII, but Edward came down with appendicitis and the coronation had to be postponed. Henry refused to pay the balance of the rent and the court held that he was excused from doing so because his purpose in renting had been frustrated by the postponement, a contingency outside the knowledge, or power to influence, of either party. The question was, to which party did the contract (implicitly) allocate the risk? Surely Henry had not intended to insure Krell against the possibility of the coronation’s being postponed, since Krell could always relet the room, at the premium rental, for the coronation’s new date. So Henry was excused.

NIPSCO is the buyer in the present case, and its defense is more properly frustration than impracticability; but the judge held that frustration is not a contract defense under the law of Indiana. He relied on an Indiana Appellate Court decision which indeed so states, Ross Clinic, Inc. v. Tabion, 419 N.E.2d 219, 223 (Ind.App.1981), but solely on the basis of an old decision of the Indiana Supreme Court, Krause v. Board of Trustees, 162 Ind. 278, 283-84, 70 N.E. 264, 265 (1904), that doesn’t even discuss the defense of frustration and anyway precedes by years the recognition of the defense by American courts. At all events, the facts of the present case do not bring it within the scope of the frustration doctrine, so we need not decide whether the Indiana Supreme Court would embrace the doctrine in a suitable case.

For the same reason we need not decide whether a force majeure clause should be deemed a relinquishment of a party’s right to argue impracticability or frustration, on the theory that such a clause represents the integrated expression of the parties’ desires with respect to excuses based on supervening events; or whether such a clause either in general or as specifically worded in this case covers any different ground from these defenses; or whether a buyer can urge impracticability under section 2-615 of the Uniform Commercial Code, which applies to this suit. Regarding the last of these questions, although the text says “seller,” Official Comment 9 to the section says that in some circumstances “the reason of the present section may well apply and entitle the buyer to the exemption,” and many courts have done just that. See, e.g., Nora Springs Coop. Co. v. Brandau, 247 N.W.2d 744 (Iowa 1976); Lawrance v. Elmore Bean Warehouse, Inc., 108 Idaho 892, 894, 702 P.2d 930, 932 (Idaho App.1985); Northern Illinois Gas Co. v. Energy Coop., Inc., supra, 122 Ill.App.3d at 954, 78 Ill.Dec. at 226, 461 N.E.2d at 1060. The rub is that Indiana has not adopted the “Official Comments” to the UCC. It has its own official comments, and they seem critical of Official Comment 9: “Comment 9 discusses ‘exemption’ for the buyer, but the text of the section is applicable only to sellers.” Burns Ind.Stat.Ann. § 26-1-2-615, Ind. Comment. It may be, therefore, that buyers cannot use section 2-615 in Indiana. But it is not clear that this has substantive significance. Section 1-103 of the Uniform Commercial Code authorizes the courts to *278apply common law doctrines to the extent consistent with the Code — this is the basis on which NIPSCO is able to plead frustration as an alternative defense to section 2-615; and the essential elements of frustration and of impracticability are the same. With section 2-615 compare Restatement, supra, §§ 261 (impossibility/impracticability) and 265 (frustration); and see id., § 265, comment a. NIPSCO gains nothing by pleading section 2-615 of the Uniform Commercial Code as well as common law frustration, and thus loses nothing by a ruling that buyers in Indiana cannot use section 2-615.

Whether or not Indiana recognizes the doctrine of frustration, and whether or not a buyer can ever assert the defense of impracticability under section 2-615 of the Uniform Commercial Code, these doctrines, so closely related to each other and to force majeure as well, see International Minerals & Chemical Corp. v. Llano, Inc., 770 F.2d 879, 885-87 (10th Cir.1985), cannot help NIPSCO. All are doctrines for shifting risk to the party better able to bear it, either because he is in a better position to prevent the risk from materializing or because he can better reduce the disutility of the risk (as by insuring) if the risk does occur. Suppose a grower agrees before the growing season to sell his crop to a grain elevator, and the crop is destroyed by blight and the grain elevator sues. Discharge is ordinarily allowed in such cases. See, e.g., Matousek v. Galligan, 104 Neb. 731, 178 N.W. 510 (1920); Pearce-Young-Angel Co. v. Charles R. Allen, Inc., 213 S.C. 578, 50 S.E.2d 698 (1948); cf. Olbum v. Old Home Manor, Inc., 313 Pa.Super. 99, 459 A.2d 757 (1983). The grower has every incentive to avoid the blight; so if it occurs, it probably could not have been prevented; and the grain elevator, which buys from a variety of growers not all of whom will be hit by blight in the same growing season, is in a better position to buffer the risk of blight than the grower is.

Since impossibility and related doctrines are devices for shifting risk in accordance with the parties’ presumed intentions, which are to minimize the costs of contract performance, one of which is the disutility created by risk, they have no place when the contract explicitly assigns a particular risk to one party or the other. As we have already noted, a fixed-price contract is an explicit assignment of the risk of market price increases to the seller and the risk of market price decreases to the buyer, and the assignment of the latter risk to the buyer is even clearer where, as in this case, the contract places a floor under price but allows for escalation. If, as is also the case here, the buyer forecasts the market incorrectly and therefore finds himself locked into a disadvantageous contract, he has only himself to blame and so cannot shift the risk back to the seller by invoking impossibility or related doctrines. See Farnsworth, supra, at 680 and n. 18; White & Summers, Handbook of the Law Under the Uniform Commercial Code 133 (2d ed. 1980). It does not matter that it is an act of government that may have made the contract less advantageous to one party. See, e.g., Connick v. Teachers Ins. & Annuity Ass’n, 784 F.2d 1018, 1022 (9th Cir.1986); Waegemann v. Montgomery Ward & Co., 713 F.2d 452, 454 (9th Cir.1983). Government these days is a pervasive factor in the economy and among the risks that a fixed-price contract allocates between the parties is that of a price change induced by one of government’s manifold interventions in the economy. Since “the very purpose of a fixed price agreement is to place the risk of increased costs on the promisor (and the risk of decreased costs on the promisee),” the fact that costs decrease steeply (which is in effect what happened here — the cost of generating electricity turned out to be lower than NIPSCO thought when it signed the fixed-price contract with Carbon County) cannot allow the buyer to walk away from the contract. In re Westinghouse Electric Corp. Uranium Contracts Litigation, 517 F.Supp. 440, 453 (E.D.Va.1981); cf. Neal-Cooper Grain Co. v. Texas Gulf Sulphur Co., 508 F.2d 283, 293 (7th Cir.1974).

*2795. This completes our consideration of NIPSCO’s attack on the damages judgment and we turn to Carbon County’s cross-appeal, which seeks specific performance in lieu of the damages it got. Carbon County’s counsel virtually abandoned the cross-appeal at oral argument, noting that the mine was closed and could not be reopened immediately — so that if specific performance (i.e., NIPSCO’s resuming taking the coal) was ordered, Carbon County would not be able to resume its obligations under the contract without some grace period. In any event the request for specific performance has no merit. Like other equitable remedies, specific performance is available only if damages are not an adequate remedy, Farnsworth, supra, § 12.6, and there is no reason to suppose them inadequate here. The loss to Carbon County from the breach of contract is simply the difference between (1) the contract price (as escalated over the life of the contract in accordance with the contract’s escalator provisions) times quantity, and (2) the cost of mining the coal over the life of the contract. Carbon County does not even argue that $181 million is not a reasonable estimate of the present value of the difference. Its complaint is that although the money will make the owners of Carbon County whole it will do nothing for the miners who have lost their jobs because the mine is closed and the satellite businesses that have closed for the same reason.. Only specific performance will help them.

But since they are not parties to the contract their losses are irrelevant. Indeed, specific performance would be improper as well as unnecessary here, because it would force the continuation of production that has become uneconomical. Cf. Farnsworth, supra, at 817-18. No one wants coal from Carbon County’s mine. With the collapse of oil prices, which has depressed the price of substitute fuels as well, this coal costs far more to get out of the ground than it is worth in the market. Continuing to produce it, under compulsion of an order for specific performance, would impose costs on society greater than the benefits. NIPSCO’s breach, though it gave Carbon County a right to damages, was an efficient breach in the sense that it brought to a halt a production process that was no longer cost-justified. See Lake River Corp. v. Carborundum Co., 769 F.2d 1284, 1289 (7th Cir.1985); Thyssen, Inc. v. S.S. Fortune Star, 777 F.2d 57, 63 (2d Cir.1985) (Friendly, J.). The reason why NIPSCO must pay Carbon County’s loss is not that it should have continued buying coal it didn’t need but that the contract assigned to NIPSCO the risk of market changes that made continued deliveries uneconomical. The judgment for damages is the method by which that risk is being fixed on NIPSCO in accordance with its undertakings.

With continued production uneconomical, it is unlikely that an order of specific performance, if made, would ever actually be implemented. If, as a finding that the breach was efficient implies, the cost of a substitute supply (whether of coal, or of electricity) to NIPSCO is less than the cost of producing coal from Carbon County’s mine, NIPSCO and Carbon County can both be made better off by negotiating a cancellation of the contract and with it a dissolution of the order of specific performance. Suppose, by way of example, that Carbon County’s coal costs $20 a ton to produce, that the contract price is $40, and that NIPSCO can buy coal elsewhere for $10. Then Carbon County would be making a profit of only $20 on each ton it sold to NIPSCO ($40-$20), while NIPSCO would be losing $30 on each ton it bought from Carbon County ($40-$10). Hence by offering Carbon County more than contract damages (i.e., more than Carbon County’s lost profits), NIPSCO could induce Carbon County to discharge the contract and release NIPSCO to buy cheaper coal. For example, at $25, both parties would be better off than under specific performance, where Carbon County gains only $20 but NIPSCO loses $30. Probably, therefore, Carbon County is seeking specific performance in order to have bargaining leverage with NIPSCO, and we can think of no reason why the law should give it such *280leverage. We add that if Carbon County obtained and enforced an order for specific performance this would mean that society was spending $20 (in our hypothetical example) to produce coal that could be gotten elsewhere for $10 — a waste of scarce resources.

As for possible hardships to workers and merchants in Hanna, Wyoming, where Carbon County’s coal mine is located, we point out that none of these people were parties to the contract with NIPSCO or third-party beneficiaries. They have no legal interest in the contract. Cf. Local 1330, United Steel Workers of America v. United States Steel Corp., 631 F.2d 1264, 1279-82 (6th Cir.1980); Serrano v. Jones & Laughlin Steel Co., 790 F.2d 1279, 1289 (6th Cir.1986). Of course the- consequences to third parties of granting an injunctive remedy, such as specific performance, must be considered, and in some cases may require that the remedy be withheld. See Weinberger v. Romero-Barcelo, 456 U.S. 305, 312-13, 102 S.Ct. 1798, 1803, 72 L.Ed.2d 91 (1982); Shondel v. McDermott, 775 F.2d 859, 868 (7th Cir.1985); Duran v. Elrod, 760 F.2d 756, 759 (7th Cir.1985); Donovan v. Robbins, 752 F.2d 1170, 1176 (7th Cir.1985). The frequent references to “public interest” as a factor in the grant or denial of a preliminary injunction invariably are references to third-party effects. See, e.g., Punnett v. Carter, 621 F.2d 578, 587-88 (3d Cir.1980). But even though the formal statement of the judicial obligation to consider such effects extends to orders denying as well as granting injunctive relief, see, e.g., Kershner v. Mazurkiewicz, 670 F.2d 440, 443 (3d Cir.1982) (en banc), the actuality is somewhat different: when the question is whether third parties would be injured by an order denying an injunction, always they are persons having a legally recognized interest in the lawsuit, so that the issue really is the adequacy of relief if the injunction is denied. In Mississippi Power & Light Co. v. United Gas Pipe Line Co., 760 F.2d 618 (5th Cir.1985), for example, a public utility sought a preliminary injunction against alleged overcharges by a supplier. If the injunction was denied and later the utility got damages, its customers would be entitled to refunds; but for a variety of reasons explained in the opinion, refunds would not fully protect the customers’ interests. The customers were the real parties in interest on the plaintiff side of the case, and their interests had therefore to be taken into account in deciding whether there would be irreparable harm (and how much) if the preliminary injunction was denied. See id. at 623-26. Carbon County does not stand in a representative relation to the workers and businesses of Hanna, Wyoming. Treating them as real parties in interest would evade the limitations on the concept of a third-party beneficiary and would place the promisor under obligations potentially far heavier than it had thought it was accepting when it signed the contract. Indeed, if we are right that an order of specific performance would probably not be carried out — that instead NIPSCO would pay an additional sum of money to Carbon County for an agreement not to enforce the order — it becomes transparent that granting specific performance would make NIPSCO liable in money damages for harms to non-parties to the contract, and it did not assume such liability by signing the contract. Cf. H.R. Moch Co. v. Rensselaer Water Co., 247 N.Y. 160, 159 N.E. 896 (1928).

Moreover, the workers and merchants in Hanna assumed the risk that the coal mine would have to close down if it turned out to be uneconomical. The contract with NIP-SCO did not guarantee that the mine would operate throughout the life of the contract but only protected the owners of Carbon County against the financial consequences to them of a breach. As Carbon County itself emphasizes in its brief, the contract was a product of the international oil cartel, which by forcing up the price of substitute fuels such as coal made costly coal-mining operations economically attractive. The OPEC cartel is not a source of vested rights to produce substitute fuels at inflated prices.

6. The last issue is the judge’s refusal to force NIPSCO to post a $181 million *281bond as a condition of obtaining a stay of execution of the damage judgment pending NIPSCO’s appeal. The issue is not moot, because our, decision upholding the judgment will not become final until NIPSCO has exhausted all its rights of further review, and meanwhile Carbon County will continue to incur whatever risk is created by the absence of a bond.

There are two grounds for rejecting Carbon County’s appeal. The first is the cross-appeal, which seeks to substitute specific performance for the damage judgment. A party that it trying to obtain specific performance in lieu of damages cannot at the same time attempt to execute a damage judgment. Bronson v. La Crosse & Milwaukee R.R., 68 U.S. 405, 409-10, 1 Wall. 405, 409-10, 17 L.Ed. 616 (1863); Price v. Franklin Investment Co., 574 F.2d 594, 597 (D.C.Cir.1978). Yet the reason for an appeal bond is to give the plaintiff security while execution is postponed. If he does not want to execute the damage judgment, because he hopes for something better, he gives up nothing by waiting till the appeals are over with before executing the damage judgment— there is no quid for the quo of a bond paid for by the defendant-appellant. Although supersedeas bonds have sometimes been granted in such cases, see New York v. Shore Realty Corp., 763 F.2d 49, 51 (2d Cir.1985); Mid-Jersey Nat’l Bank v. Fidelity-Mortgage Investors, 518 F.2d 640, 642 (3d Cir.1975); Knapp v. Kinsey, 249 F.2d 797, 800 (6th Cir.1957), none of the cases discusses the propriety of doing so. Lacking omniscience, we hesitate to say that requiring a bond would never be proper in such a case, but there would have to be a good reason for it and Carbon County has provided none.

Second, as explained in Olympia Equipment Leasing Co. v. Western Union Telegraph Co., 786 F.2d 794 (7th Cir.1986), it is a misreading of Rule 62(d) of the Federal Rules of Civil Procedure to suggest that an appellant who wants to stay execution pending appeal must post a bond. The rule requires him to post a bond if he wants an automatic stay, but not if he is content to throw himself on the district judge’s discretion. See also Lightfoot v. Walker, 797 F.2d 505, 506 (7th Cir.1986). An appeal bond usually costs one percent of the amount secured, which in the case of a $181 million judgment is almost $2 million. That is not small change; and if the district judge is satisfied that the expenditure is unnecessary to protect the appellee, he does not have to insist that it be spent. NIPSCO has assets of more than $4 billion, revenues of almost $2 billion a year, and a net worth of more than $1 billion. A public utility, it is in no financial jeopardy, it is not about to place its assets beyond the reach of this judgment creditor, and it is, in short, good for the $181 million. Should NIPSCO’s ability to pay begin to deteriorate, Carbon County can always petition the district judge for supplementary relief; he has even required periodic reports from NIPSCO to make it easier to monitor the company’s financial health.

To summarize, the appeal from the grant of the preliminary injunction is dismissed as moot; the other orders appealed from are affirmed. No costs will be awarded in this court, since we have turned down Carbon County’s appeals as well as NIPSCO’s.

So Ordered.

POSNER, Circuit Judge.

This appeal from the grant of a permanent injunction raises fundamental issues concerning the propriety of injunctive relief 775 F.Supp. 1192 (E.D.Wis.1991). The essential facts are simple. Walgreen has operated a pharmacy in the Southgate Mall in Milwaukee since its opening in 1951. Its current lease, signed in 1971 and carrying a 30-year, 6-month term, contains, as had the only previous lease, a clause in which the landlord, Sara Creek, promises not to lease space in the mall to anyone else who wants to operate a pharmacy or a store containing a pharmacy. Such an exclusivity clause, common in shopping-center leases, is occasionally challenged on antitrust grounds, Milton Handler & Daniel E. Laza-roff, “Restraint of Trade and the Restatement (Second) of Contracts,” 57 N.Y.U.L.Rev. 669, 683-708 (1982); Note, “The Antitrust Implications of Restrictive Covenants in Shopping Center Leases,” 86 Harv.L.Rev. 1201 (1973) — implausibly enough, given the competition among malls; but that is an issue for another day, since in this appeal Sara Creek does not press the objection it made below to the clause on antitrust grounds.

In 1990, fearful that its largest tenant— what in real estate parlance is called the “anchor tenant” — having gone broke was about to close its store, Sara Creek informed Walgreen that it intended to buy out the anchor tenant and install in its place a discount store operated by Phar-Mor Corporation, a “deep discount” chain, rather than, like Walgreen, just a “discount” chain. Phar-Mor's store would occupy 100,000 square feet, of which 12,000 would be occupied by a pharmacy the same size as Walgreen’s. The entrances to the two stores would be within a couple of hundred feet of each other.

Walgreen filed this diversity suit for breach of contract against Sara Creek and Phar-Mor and asked for an injunction against Sara Creek’s letting the anchor premises to Phar-Mor. After an evidentia-ry hearing, the judge found a breach of Walgreen’s lease and entered a permanent injunction against Sara Creek’s letting the anchor tenant premises to Phar-Mor until the expiration of Walgreen’s lease. He did this over the defendants’ objection that Walgreen had failed to show that its remedy at law — damages—for the breach of the exclusivity clause was inadequate. Sara Creek had put on an expert witness who testified that Walgreen’s damages could be readily estimated, and Walgreen had countered with evidence from its employees that its damages would be very difficult to compute, among other reasons because they included intangibles such as loss of goodwill.

Sara Creek reminds us that damages are the norm in breach of contract as in other cases. Many breaches, it points out, are “efficient” in the sense that they allow resources to be moved into a more valuable use. Patton v. Mid-Continent Systems, Inc., 841 F.2d 742, 750-51 (7th Cir.1988). Perhaps this is one — the value of Phar-Mor’s occupancy of the anchor premises may exceed the cost to Walgreen of facing increased competition. If so, society will be better off if Walgreen is paid its damages, equal to that cost, and Phar-Mor is allowed to move in rather than being kept out by an injunction. That is why injunctions are not granted as a matter of course, but only when the plaintiff’s damages remedy is inadequate. Northern Indiana Public Service Co. v. Carbon County Coal Co., 799 F.2d 265, 279 (7th Cir.1986). Walgreen’s is not, Sara Creek argues; the projection of business losses due to increased competition is a routine exercise in calculation. Damages representing either the present value of lost future profits or (what should be the equivalent, Carusos v. Briarcliff, Inc., 76 Ga. App. 346, 351-52, 45 S.E.2d 802, 806-07 *275(1947)) the diminution in the value of the leasehold have either been awarded or deemed the proper remedy in a number of reported cases for breach of an exclusivity clause in a shopping-center lease. Coach House of Ward Parkway, Inc. v. Ward Parkway Shops, Inc., 471 S.W.2d 464, 473 (Mo.1971); Krikorian v. Dailey, 171 Va. 16, 197 S.E. 442 (1938); PNY Realty Corp. v. Chong Leung Restaurant, 116 Misc.2d 1035, 457 N.Y.S.2d 358 (1982); Saucier v. John-Clai Co., 408 So.2d 27 (La.App.1981); Barr & Sons, Inc. v. Cherry Hill Center, Inc., 90 N.J.Super. 358, 376, 217 A.2d 631, 641 (1966); Renee Cleaners, Inc. v. Good Deal Super Markets of N.J., Inc., 89 N.J.Super. 186, 214 A.2d 437 (App.Div. 1965); Carusos v. Briarcliff, Inc., supra; Annot., “Validity, Construction, and Effect of Lessor’s Covenant Against Use of His Other Property in Competition with the Lessee-Covenantee,” 97 A.L.R.2d 4, 111-117 (§ 28) (1964 and Supp.1983). Why, Sara Creek asks, should they not be adequate here?

Sara Creek makes a beguiling argument that contains much truth, but we do not think it should carry the day. For if, as just noted, damages have been awarded in some cases of breach of an exclusivity clause in a shopping-center lease, injunctions have been issued in others. Handy Andy Home Improvement Centers, Inc. v. American National Bank & Trust Co., 177 Ill.App.3d 647, 126 Ill.Dec. 852, 532 N.E.2d 537 (1988); De Koven Drug Co. v. First National Bank, 27 Ill.App.3d 798, 800, 327 N.E.2d 378, 379 (1975); Regis Corp. v. Fusco Corp., 496 So.2d 833, 835 (Fla.App.1986); Belvin v. Sikes, 2 So.2d 65 (La.App.1941); Child World, Inc. v. South Towne Centre, Ltd., 634 F.Supp. 1121, 1134-35 (S.D.Ohio 1986). The choice between remedies requires a balancing of the costs and benefits of the alternatives. Hecht Co. v. Bowles, 321 U.S. 321, 329, 64 S.Ct. 587, 591, 88 L.Ed. 754 (1944); Yakus v. United States, 321 U.S. 414, 440, 64 S.Ct. 660, 674, 88 L.Ed. 834 (1944). The task of striking the balance is for the trial judge, subject to deferential appellate review in recognition of its particularistic, judgmental, fact-bound character. K-Mart Corp. v. Oriental Plaza, Inc., 875 F.2d 907, 915 (1st Cir.1989). As we said in an appeal from a grant of a preliminary injunction — but the point is applicable to review of a permanent injunction as well— “The question foi us [appellate judges] is whether the [district] judge exceeded the bounds of permissible choice in the circumstances, not what we would have done if we had been in his shoes.” Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380, 390 (7th Cir.1984).

The plaintiff who seeks an injunction has the burden of persuasion — damages are the norm, so the plaintiff must show why his case is abnormal. But when, as in this case, the issue is whether to grant a permanent injunction, not whether to grant a temporary one, the burden is to show that damages are inadequate, not that the denial of the injunction will work irreparable harm. “Irreparable” in the injunction context means not rectifiable by the entry of a final judgment. Diginet, Inc. v. Western Union ATS, Inc., 958 F.2d 1388, 1393 (7th Cir.1992); Vogel v. American Society of Appraisers, 744 F.2d 598, 599 (7th Cir.1984). It has nothing to do with whether to grant a permanent injunction, which, in the usual case anyway, is the final judgment. The use of “irreparable harm” or “irreparable injury” as synonyms for inádequate remedy at law is a confusing usage. It should be avoided. Owen M. Fiss & Doug Rendleman, Injunctions 59 (2d ed. 1984).

The benefits of substituting an injunction for damages are twofold. First, it shifts the burden of determining the cost of the defendant’s conduct from the court to the parties. If it is true that Walgreen’s damages are smaller than the gain to Sara Creek from allowing a second pharmacy into the shopping mall, then there must be a price for dissolving the injunction that will make both parties better off. Thus, the effect of upholding the injunction would be to substitute for the costly processes of forensic fact determination the less costly processes of private negotiation. Second, a premise of our free-market system, and the lesson of experience here and *276abroad as well, is that prices and costs are more accurately determined by the market than by government. A battle of experts is a less reliable method of determining the actual cost to Walgreen of facing new competition than negotiations between Walgreen and Sara Creek over the price at which Walgreen would feel adequately compensated for having to face that competition.

That is the benefit side of injunctive relief but there is a cost side as well. Many injunctions require continuing supervision by the court, and that is costly. Roland Machinery Co. v. Dresser Industries, Inc., supra, 749 F.2d at 391-92; Rodriguez v. VIA Metropolitan Transit System, 802 F.2d 126, 132 (5th Cir.1986); Bethlehem Engineering Export Co. v. Christie, 105 F,2d 933, 935 (2d Cir.1939) (L. Hand, J.). A request for specific performance (a form of mandatory injunction) of a franchise agreement was refused oh this ground in North American Financial Group, Ltd. v. S.M.R. Enterprises, Inc., 583 F.Supp. 691, 699 (N.D.Ill.1984); see Edward Yorio, Contract Enforcement: Specific Performance and Injunctions § 3.3.2 (1989). This ground was also stressed in Rental Development Corp. v. Lavery, 304 F.2d 839, 841-42 (9th Cir.1962), a case involving a lease. Some injunctions are problematic because they impose costs on third parties. Shondel v. McDermott, 775 F.2d 859, 868 (7th Cir.1985). A more subtle cost of in-junctive relief arises from the situation that economists call “bilateral monopoly,” in which two parties can deal only with each other: the situation that an injunction creates. Goldstick v. I.C.M. Realty, 788 F.2d 456, 463 (7th Cir.1986); Milbrew, Inc. v. Commissioner, 710 F.2d 1302, 1306-07 (7th Cir.1983); Chicago & North Western Transportation Co. v. United States, 678 F.2d 665, 667 (7th Cir.1982). The sole seller of widgets selling to the sole buyer of that product would be an example. But so will be the situation confronting Walgreen and Sara Creek if the injunction is upheld. Walgreen can “sell” its injunctive right only to Sara Creek, and Sara Creek can “buy” Walgreen’s surrender of its right to enjoin the leasing of the anchor tenant’s space to Phar-Mor only from Walgreen. The lack of alternatives in bilateral monopoly creates a bargaining range, and the costs of negotiating to a point within that range may be high. Suppose the cost to Walgreen of facing the competition of Phar-Mor at the Southgate Mall would be $1 million, and the benefit to Sara Creek of leasing to Phar-Mor would be $2 million. Then at any price between those figures for a waiver of Walgreen’s injunctive right both parties would be better off, and we expect parties to bargain around a judicial assignment of legal rights if the assignment is inefficient. R.H. Coase, “The Problem of Social Cost,” 3 J. Law & Econ. 1 (1960). But each of the parties would like to engross as much of the bargaining range as possible — Walgreen to press the price toward $2 million, Sara Creek to depress it toward $1 million. With so much at stake, both parties will have an incentive to devote substantial resources of time and money to the negotiation process. The process may even break down, if one or both parties want to create for future use a reputation as a hard bargainer; and if it does break down, the injunction will have brought about an inefficient result. All these are in one form or another costs of the injunctive process that can be avoided by substituting damages.

The costs and benefits of the damages remedy are the mirror of those of the in-junctive remedy. The damages remedy avoids the cost of continuing supervision and third-party effects, and the cost of bilateral monopoly as well. It imposes costs of its own, however, in the form of diminished accuracy in the determination of value, on the one hand, and of the parties’ expenditures on preparing and presenting evidence of damages, and the time of the court in evaluating the evidence, on the other.

The weighing up of all these costs and benefits is the analytical procedure that is or at least should be employed by a judge asked to enter a permanent injunction, with the understanding that if the balance is even the injunction should be withheld. The judge is not required to explicate every *277detail of the analysis and he did not do so here, but as long we are satisfied that his approach is broadly consistent with a proper analysis we shall affirm; and we are satisfied here. The determination of Walgreen’s damages would have been costly in forensic resources and inescapably inaccurate. Roland Machinery Co. v. Dresser Industries, Inc., supra, 749 F.2d at 386; Olympia Equipment Leasing Co. v. Western Union Telegraph Co., 797 F.2d 370, 382 (7th Cir.1986); K-Mart Corp. v. Oriental Plaza, Inc., supra, 875 F.2d at 914-16. The lease had ten years to run. So Walgreen would have had to project its sales revenues and costs over the next ten years, and then project the impact on those figures of Phar-Mor’s competition, and then discount that impact to present value. All but the last step would have been fraught with uncertainty.

We may have given too little weight to such uncertainties in American Dairy Queen Corp. v. Brown-Port Co., 621 F.2d 255, 257 n. 2 (7th Cir.1980), but in that case the district judge had found that the remedy at law was adequate in the circumstances and the movant had failed to make its best argument for inadequacy in the district court. Id. at 259. It is difficult to forecast the profitability of a retail store over a decade, let alone to assess the impact of a particular competitor on that profitability over that period. Of course one can hire an expert to make such predictions, Glen A. Stankee, “Econometric Forecasting of Lost Profits: Using High Technology to Compute Commercial Damages,” 61 Fla.B.J. 83 (1987), and if injunctive relief is infeasible the expert’s testimony may provide a tolerable basis for an award of damages. We cited cases in which damages have been awarded for the breach of an exclusivity clause in a shopping-center lease. But they are awarded in such circumstances not because anyone thinks them a clairvoyant forecast but because it is better to give a wronged person a crude remedy than none at all. It is the same theory on which damages are awarded for a disfiguring injury. No one thinks such injuries readily monetizable, City of Panama, 101 U.S. 453, 464, 25 L.Ed. 1061 (1880); McCarty v. Pheasant Run, Inc., 826 F.2d 1554, 1557 (7th Cir.1987); Marcus L. Plant, “Damages for Pain and Suffering,” 19 Ohio St.L.J. 200, 205-06 (1958), but a crude estimate is better than letting the wrongdoer get off scot-free (which, not incidentally, would encourage more such injuries). Randall R. Bovbjerg et al, “Valuing Life and Limb in Tort: Scheduling ‘Pain and Suffering,’ ” 83 Nw.U.L.Rev. 908 (1989). Sara Creek presented evidence of what happened (very little) to Walgreen when Phar-Mor moved into other shopping malls in which Walgreen has a pharmacy, and it. was on the right track in putting in comparative evidence. But there was a serious question whether the other malls were actually comparable to the Southgate Mall, so we cannot conclude, in the face of the district judge’s contrary conclusion, that the existence of comparative evidence dissolved the difficulties of computing damages in this case. Sara Creek complains that the judge refused to compel Walgreen to produce all the data that Sara Creek needed to demonstrate the feasibility of forecasting Walgreen’s damages. Walgreen resisted, on grounds of the confidentiality of the data and the cost of producing the massive data that Sara Creek sought. Those are legitimate grounds; and the cost (broadly conceived) they expose of pretrial discovery, in turn presaging complexity at trial, is itself a cost of the damages remedy that injunctive relief saves.

Damages are not always costly to compute, or difficult to compute accurately. In the standard case of a seller’s breach of a contract for the sale of goods where the buyer covers by purchasing the same product in the market, damages are readily calculable by subtracting the contract price from the market price and multiplying by the quantity specified in the contract. But this is not such a case and here damages would be a costly and inaccurate remedy; and on the other side of the balance some of the costs of an injunction are absent and the cost that is present seems low. The injunction here, like one enforcing a covenant not to compete (standardly enforced by injunction, Yorio, supra, 401-08), is a *278simple negative injunction — Sara Creek is not to lease space in the Southgate Mall to Phar-Mor during the term of Walgreen’s lease — and the costs of judicial supervision and enforcement should be negligible. There is no contention that the injunction will harm an unrepresented third party. It may harm Phar-Mor but that harm will be reflected in Sara Creek’s offer to Walgreen to dissolve the injunction. (Anyway Phar-Mor is a party.) The injunction may also, it is true, harm potential customers of Phar-Mor — people who would prefer to shop at a deep-discount store than an ordinary discount store — but their preferences, too, are registered indirectly. The more business Phar-Mor would have, the more rent it will be willing to pay Sara Creek, and therefore the more Sara Creek will be willing to pay Walgreen to dissolve the injunction.

The only substantial cost of the injunction in this case is that it may set off a round of negotiations between the parties. In some cases, illustrated by Boomer v. Atlantic Cement Co., 26 N.Y.2d 219, 309 N.Y.S.2d 312, 257 N.E.2d 870 (1970), this consideration alone would be enough to warrant the denial of injunctive relief. The defendant’s factory was emitting cement dust that caused the plaintiffs harm monetized at less than $200,000, and the only way to abate the harm would have been to close down the factory, which had cost $45 million to build. An injunction against the nuisance could therefore have created a huge bargaining range (could, not would, because it is unclear what the current value of the factory was), and the costs of negotiating to a point within it might have been immense. If the market value of the factory was actually $45 million, the plaintiffs would be tempted to hold out for a price to dissolve the injunction in the tens of millions and the factory would be tempted to refuse to pay anything more than a few hundred thousand dollars. Negotiations would be unlikely to break down completely, given such a bargaining range, but they might well be protracted and costly. There is nothing so dramatic here. Sara Creek does not argue that it will have to close the mall if enjoined from leasing to Phar-Mor. Phar-Mor is not the only potential anchor tenant. Liza Danielle, Inc. v. Jamko, Inc., 408 So.2d 735, 740 (Fla.App. 1982), on which Sara Creek relies, presented the converse case where the grant of the injunction would have forced an existing tenant to close its store. The size of the bargaining range was also a factor in the denial of injunctive relief in Gitlitz v. Plankinton Building Properties, Inc., 228 Wis. 334, 339-40, 280 N.W. 415, 418 (1938).

To summarize, the judge did not exceed the bounds of reasonable judgment in concluding that the costs (including forgone benefits) of the damages remedy would exceed the costs (including forgone benefits) of an injunction. We need not consider whether, as intimated by Walgreen, exclusivity clauses in shopping-center leases should be considered presumptively enforceable by injunctions. Although we have described the choice between legal and equitable remedies as one for case-by-case determination, the courts have sometimes picked out categories of case in which injunctive relief is made the norm. The best-known example is specific performance of contracts for the sale of real property. Anderson v. Onsager, 155 Wis.2d 504, 455 N.W.2d 885 (1990); Okaw Drainage District v. National Distillers & Chemical Corp., 882 F.2d 1241, 1248 (7th Cir.1989); Anthony T. Kronman, “Specific Performance,” 45 U.Chi.L.Rev. 351, 355 and n. 20 (1978). The rule that specific performance will be ordered in such cases as a matter of course is a generalization of the considerations discussed above. Because of the absence of a fully liquid market in real property and the frequent presence of subjective values (many a homeowner, for example, would not sell his house for its market value), the calculation of damages is difficult; and since an order of specific performance to convey a piece of property does not create a continuing relation between the parties, the costs of supervision and enforcement if specific performance is ordered are slight. The exclusivity clause in Walgreen’s lease relates to real estate, but we hesitate to suggest that every contract involving real estate should *279be enforceable as a matter of course by injunctions. Suppose Sara Creek had covenanted to keep the entrance to Walgreen’s store free of ice and snow, and breached the covenant. An injunction would require continuing supervision, and it would be easy enough if the injunction were denied for Walgreen to hire its own ice and snow remover and charge the cost to Sara Creek. Cf. City of Michigan City v. Lake Air Corp., 459 N.E.2d 760 (Ind.App.1984). On the other hand, injunctions to enforce exclusivity clauses are quite likely to be justifiable by just the considerations present here — damages are difficult to estimate with any accuracy and the injunction is a one-shot remedy requiring no continuing judicial involvement. So there is an argument for making injunctive relief presumptively appropriate in such eases, but we need not decide in this ease how strong an argument.

AFFIRMED.

HARLINGTON WOOD, Jr., Senior Circuit Judge,

concurring.

I gladly join in the affirmance reached in Judge Posner’s expert analysis.

VAN OOSTERHOUT, Circuit Judge.

Plaintiff Southwest Engineering Company, hereinafter called Southwest, has appealed from summary judgment dismissing its complaint against the United States for recovery of $8,300 withheld as liquidated damages for delay in performance on four construction contracts entered into between Southwest and the United States.

This appeal is before us upon an agreed statement of the record. Four contracts entered into between Southwest and the Government called for the construction by Southwest of three V. O. R. radio facilities at Readsville, Blackwater, and Maryland Heights, Missouri, and for a high intensity approach light lane at Lambert Field, Missouri. Each of the contracts fixed a completion date and provided for liquidated damages on a per diem basis for each day’s delay beyond the agreed completion date. The agreed liquidated damage on the Lambert Field project was $100 per day, and $50 per day on each of the other projects. Each contract contained the general provisions set forth in Standard Form 23-A (March 1953) prescribed by the General Services Administration, including, among others, a provision that plaintiff was not to be charged with delays “due to unforeseeable causes beyond the control and without the fault or negligence of the Contractor, including, but not restricted to, acts of God, or of the public enemy, acts of the Government, in either its sovereign or contractual capacity, acts of another contractor in the performance of a contract with the Government, fires, floods, epidemics, quarantine restrictions, strikes, freight embargoes, and unusually severe weather, or delays of subcontractors or suppliers due to such causes.” Within ten days from the beginning of any such delay, plaintiff was to notify the contracting officer in writing of the causes of delay. The contracting officer was to ascertain the facts and extent of the delay, and extend the time for completing the work when “in his judgment the findings of fact justify such an extension.” His findings of fact were to be final, subject only to appeal to the “head of the department.”

The Blackwater project was completed 97 days late. Plaintiff requested time extensions, alleging the delay resulted from causes for which it was not responsible, including acts and omissions of defendant. Administrative appeals resulted in extensions by the C.A.A. because of delays by the Government, and, later on, a remission of an additional $4,200 liquidated damages by the Comptroller General under the authority of 41 U.S.C. § 256a, because of late delivery of Government-furnished material. This left $550 (11 days) withheld as liquidated damages on the Blackwater project.

The Readsville project was completed 84 days late. Administrative appeals resulted in a three-days extension, leaving liquidated damages for 81 days totalling $4,050.

The Maryland Heights project was completed 48 days late. On administra*1000tive appeal, a fourteen-days extension was allowed resulting in liquidated damages of $1,700 for 34 days delay.

The Lambert Field project was completed 54 days late, 34 days extension was granted, leaving $2,000 as liquidated damages for 20 days delay.

The parties stipulated “although each project was not completed until after the date prescribed in its contract, defendant suffered no actual damage on any project.” The Government withheld the liquidated damages for delays provided in the contracts after giving credit for the extensions of time administratively allowed. This suit by Southwest is for recovery of the $8,300 so withheld as liquidated damages. The Government counterclaimed for the liquidated damages here involved. Plaintiff by reply admitted the projects were not completed within the time limits as administratively extended but denied the Government was entitled to liquidated damages because the Government caused and contributed to the delays and because the Government suffered no actual damages.

The trial court determined that the Government was entitled to liquidated damages in the amount claimed, offset such damages which equaled the amount of payments withheld, sustained the motion for summary judgment, and dismissed the complaint.

Southwest as a basis for reversal urges the trial court’s decision was induced by two erroneous views of the law, either of which requires reversal. The points relied upon for reversal are thus stated:

1. “The court erred in awarding the Government liquidated damages because the Government, in order to enforce a contract provision for liquidated damages for delay, must not cause or contribute to such delay, and if the Government does cause or contribute to such delay, it cannot assess liquidated damages.”

2. “The court erred in awarding the Government liquidated damages because a contract provision for liquidated damages is clearly a penalty and not enforceable where the party seeking to enforce it formally admits he sustained no actual damage.”

At the outset, we observe that the contracts here involved were entered into pursuant to federal law by an authorized federal agency. Federal law controls in the construction and determination of rights under federal contracts. Priebe & Sons, Inc. v. United States, 332 U.S. 407, 411, 68 S.Ct. 123, 92 L.Ed. 32; Clearfield Trust Co. v. United States, 318 U.S. 363, 63 S.Ct. 573, 87 L.Ed. 838; United States v. Le Roy Dyal Co., 3 Cir., 186 F.2d 460, 461.

In the Priebe case, the Supreme Court-states :

“It is customary, where Congress has not adopted a different standard, to apply to the construction of government contracts the principles of general contract law. United States v. Standard Rice Co., 323 U.S. 106, 111, [65 S.Ct. 145, 147, 89 L.Ed. 104] and cases cited. That has been done in other cases where the Court has considered the enforceability of ‘liquidated damages’ provisions in government contracts.” 332 U.S. 407, 411, 68 S.Ct. 123, 125.

Southwest’s first point, to the effect that the Government is not entitled to liquidated damages because it caused or contributed to the delay, is without merit. It is true that in the administrative appeals a finding was made that some of the delay was caused by the Government and that some other delays were excusable. No damages were withheld for excusable delays as administratively determined. In rejecting a similar contention, the Supreme Court in Robinson v. United States, 261 U.S. 486, 488, 43 S.Ct. 420, 421, 67 L.Ed. 760, states:

“The fact that the government’s action caused some of the delay presents no legal ground for denying it compensation for loss suffered wholly through the fault of the contractor. Since the contractor agreed to pay at a specified rate for each day’s *1001delay not caused by the government, it was clearly the intention that it should pay for some days’ delay at that rate, even if it were relieved from paying for other days, because of the government’s action. * * ”

See Wells and Wells, Inc. v. United States, 8 Cir., 269 F.2d 412, 415; Union Paving Co. v. United States, 115 F.Supp. 179, 126 Ct.Cl. 478.

United States v. United Eng’r & Contracting Co., 234 U.S. 236, 34 S.Ct. 843, 58 L.Ed. 1294, relied upon heavily by Southwest, is clearly distinguishable as pointed out by the Supreme Court in Robinson, where it said :

“The ease is wholly unlike United States v. United Engineering Co., 234 U.S. 236 [34 Sup.Ct. 843, 58 L. Ed. 1294] upon which claimant relies. The question there was one of construction. The lower court found, as a fact that, but for the government’s action, the work would have been completed within the contract period, and this court construed the provision for liquidated damages as not applicable to such case. Here the question is not properly one of construction.” 261 U.S. 486, 489, 43 S.Ct. 420, 421.

The trial court found upon review of the administrative record that the findings there made with respect to excusable delays were supported by substantial evidence and were entitled to be upheld upon the basis of the teaching of United States v. Carlo Bianchi & Co., 373 U.S. 709, 83 S.Ct. 1409, 10 L.Ed.2d 652. We agree.

Southwest’s second point in substance is that the parties’ stipulation that the Government suffered no actual damage bars any recovery of liquidated damages. Such stipulation was made. It relates to the situation as it existed at the time of the completion of the work. The stipulation does not go to the extent of agreeing that the parties at the time of contracting did not reasonably contemplate that damages would flow from a delay in performance.

The contracts prescribe the payments to be made for delays as liquidated damages. Southwest urges that such provision is a penalty provision. The agreed record is highly condensed and does not show the contract price for the various projects, but such projects appear to be substantial. There is no showing that the liquidated damages for delay provided for are beyond damages reasonably contemplated by the parties at the time of the contract.

Two requirements must be considered to determine whether the provision included in the contract fixing the amount of damages payable on breach will be interpreted as an enforceable liquidated damage clause rather than an unforceable penalty clause: First, the amount so fixed must be a reasonable forecast of just compensation for the harm that is caused by the breach, and second, the harm that is caused by the breach must be one that is incapable or very difficult of accurate estimation. See J. D. Street & Co. v. United States, 8 Cir., 256 F.2d 557, 559; United States v. Le Roy Ryal Co., supra; Restatement, Contracts § 339.

Whether these requirements have been complied with must be viewed as of the time the contract was executed rather than when the contract was breached or at some other subsequent time. Courts presently look with candor upon provisions that are deliberately entered into between parties and therefore do not look with disfavor upon liquidated damage stipulations. Rex Trailer Co. v. United States, 350 U.S. 148, 151, 76 S.Ct. 219, 100 L.Ed. 149; Priebe & Sons, Inc. v. United States supra; United States v. Bethlehem Steel Co., 205 U.S. 105, 119, 27 S.Ct. 450, 51 L.Ed. 731; Sun Printing & Publishing Ass’n v. Moore, 183 U.S. 642, 660, 22 S.Ct. 240, 46 L.Ed. 366.

In the Bethlehem Steel case, the Court sets out the standard to be applied for determining whether a liquidated damage provision can be upheld as follows: “The question always is, What did the parties intend by the language used? When such intention is ascertained it is *1002ordinarily the duty of the court to carry it out.” 205 U.S. 105, 119, 27 S.Ct. 450, 455.

Here Southwest has failed to establish that the intention of the parties was to do anything other than execute a valid liquidated damage provision,

Priebe & Sons, Inc. v. United States, supra, involved a Government contract for purchase of dried eggs for foreign shipment. The Court observes the contract contains two provisions for liquidated damages. The one involved in the case related to failure to have eggs inspected and ready for delivery for the date specified in the order. The court construed the contract to mean that performance was not due until demand for delivery and held that since inspection was made before such demand, the provision for liquidated damages could not possibly be a reasonable forecast of just compensation for breach of contract. The Court further observes that the contract contains a provision for liquidated damages for delay in delivery upon demand and with respect to such provision, breach of which was not involved, states: “It likewise is apparent that the only thing which could possibly injure the Government would be failure to get prompt performance when delivery was due. We have no doubt of the validity of the provision for ‘liquidated damages’ when applied under those circumstances.” 332 U.S. 407, 412, 68 S.Ct. 123, 126.

The Court also speaks of standards applicable to allowance of liquidated damages, stating:

“Today the law does not look with disfavor upon ‘liquidated damages’ provisions in contracts. When they are fair and reasonable attempts to fix just compensation for anticipated loss caused by breach of contract, they are enforced. * * * They serve a particularly useful function when damages are uncertain in nature or amount or are unmeasurable, as is the case in many government contracts. * * * And the fact that the damages suffered are shown to be less than the damages contracted for is not fatal. These provisions are to be judged as of the time of making the contract.” 332 U.S. 407, 411-412, 68 S.Ct. 123, 126.

In Rex Trailer Co. v. United States, supra, recovery was allowed upon the statutory penalty imposed for fraud in obtaining a Government contract. The Court states: “Liquidated-damage provisions, when reasonable, are not to be regarded as penalties, * * * ” 350 U.S. 148, 151, 76 S.Ct. 219, 221. And then says:

“The Government’s recovery here is comparable to the recovery under liquidated-damage provisions which fix compensation for anticipated loss. As this Court recognized in Priebe & Sons v. United States, 332 U.S. 407, 411-412, [68 S.Ct. 123, 126, 92 L.Ed. 32] liquidated damages ‘serve a particularly useful function when damages are uncertain in nature or amount or are unmeasurable, as is the case in many government contracts. * * * ’ And the fact that no damages are shown is not fatal.” 350 U.S. 148, 153, 76 S.Ct. 219, 222.

In United States v. J. D. Streett & Co., E.D.Mo., 151 F.Supp. 469, 472, Judge Harper, in upholding a claim for liquidated damages, states:

“Although it is clear that a finding that a provision is one for liquidated damages requires that damages could be anticipated at the time of execution of the contract, whether actual damage did or did not occur or was not proved to have occurred does not prevent recovery.”

We affirmed upon appeal. J. D. Streett & Co. v. United States, 8 Cir., 256 F.2d 557.

The late Judge Goodrich in United States v. Le Roy Dyal Co., supra, in a well-considered opinion dealing with numerous aspects of the liquidated damage issue, states:

“If a provision for liquidated damages is upheld the fact that actual damage did or did not occur or was *1003not proved to have occurred does not prevent the recovery of the stipulated sum. There is some dissent on this point, but the statement just made represents the great weight of decided cases.” 186 F.2d 460, 462.

In Frick Co. v. Rubel Corp., 2 Cir., 62 F.2d 765, 768, the late Judge Learned Hand allowed liquidated damages, stating inter alia:

“My brothers think, though I do not, that evidence as to the actual loss was not material to the issue of the losses in contemplation, though we all agree that it is the comparison of the liquidated damages with the last, not the first, which can raise the point at all.”

25 C.J.S. Damages § 115d; 15 Am. Jur., Damages, § 263, and 34 A.L.R. 1336, 1341, and supporting cases cited in such authorities, indicate that the majority view is that proof of actual damages is not required to sustain an action for liquidated damages, unless the contracts so provide, at least in situations where damages could reasonably be anticipated at the time of contracting.

We recognize that there are cases, including Massman Const. Co. v. City Council of Greenville, 5 Cir., 147 F.2d 925, Rispin v. Midnight Oil Co., 9 Cir., 291 F. 481, 34 A.L.R. 1331, and Northwest Fixture Co. v. Kilbourne & Clark Co., 9 Cir., 128 F. 256, cited by Southwest, which reach a contrary result.

We believe that the cases holding that the situation existing at the time of the contract is controlling in determining the reasonableness of liquidated damages are based upon sound reasoning and represent the weight of authority. Where parties have by their contract agreed upon a liquidated damage provision as a reasonable forecast of just compensation for breach of contract and damages are difficult to estimate accurately, such provision should be enforced. If in the course of subsequent developments, damages prove to be greater than those stipulated, the party entitled to damages is bound by the liquidated damage agreement. It is not unfair to hold the contractor performing the work to-such agreement if by reason of later developments damages prove to be less or nonexistent. Each party by entering into such contractual provision took a calculated risk and is bound by reasonable contractual provisions pertaining to liquidated damages. . j

Southwest has completely failed to demonstrate that the court committed an error of law in determining that absence of actual damages at the time of breach of the contract or thereafter does not bar recovery of liquidated damages. The court at least impliedly found that the liquidated damage provisions of the contracts here involved were reasonable when viewed in the light of circumstances existing at the time the contract was entered into. We find nothing in the record which would compel a contrary conclusion. The court committed no error in allowing liquidated damages.

The judgment appealed from is affirmed.

*303Opinion

BRUINIERS, J.

These consolidated appeals are from a judgment after trial in a consumer class action against wireless telephone carrier Sprint Spectrum, L.P. (Sprint), challenging its policy of charging early termination fees (ETF’s) to customers terminating service prior to expiration of defined contract periods.1 The trial court found the ETF’s to be unlawful penalties under Civil Code section 1671, subdivision (d),2 enjoined enforcement, and granted restitution/damages to the plaintiff class in the amount of ETF’s collected by Sprint during the class period, $73,775,975. A jury found that class members who had been charged ETF’s had violated the terms of their contracts with Sprint, and that Sprint’s actual damages exceeded the ETF charges Sprint had collected. The resulting setoff negated any monetary recovery to the class. The trial court, reasoning that the jury had failed to follow its instructions on Sprint’s actual damages, granted the plaintiffs’3 motion for a partial new trial on that issue.

Sprint appeals the decision invalidating the ETF’s and enjoining their enforcement, and the court’s grant of the motion for partial new trial on damages. Plaintiffs cross-appeal, alleging that the trial court erred in permitting Sprint to assert damage claims as setoffs to class claims for recovery of ETF’s paid. In the published portions of this opinion we address the issues of federal preemption and the application of section 1671, subdivision (d). We affirm in all respects.

I. Background

Procedural History

Sprint is a national cellular service carrier, providing cellular telephone service in California. In 2003, lawsuits were filed in Alameda County and in Orange County against Sprint and other cellular service providers4 alleging that the ETF’s violated California consumer protection laws and constituted *304unauthorized penalties under section 1671.5 These actions and others were coordinated under Judicial Council order (Code Civ. Proc., § 404.3; Cal. Rules of Court, rule 3.524) before Judge Ronald Sabraw in the Alameda County Superior Court as Cellphone Termination Fee Cases (JCCP No. 4332). (See Gatton, supra, 152 Cal.App.4th at p. 575, fn. 1.)

On June 9, 2006, Judge Ronald Sabraw, the then designated coordination judge, certified a class in the related cases defined as: “ ‘All persons who (1) had a wireless telephone personal account with [Sprint] with a California area code and a California billing addressf] who (2) cancelled the account at any time from July 23, 1999, through [March 18, 2007], and (3) were charged an early termination fee in connection with that cancellation.’ ”6 The class certification was “expressly predicated” on an “aggregate approach to monetary relief and the related setoff and cross-claim issues.” Thus, if the ETF’s were found to be illegal and unenforceable, the wireless carriers would still potentially be entitled to offset against any class recovery for their actual damages in the form of lost profits.7

Pursuant to case management orders in the coordination proceedings, the ETF claims against Sprint were separately pled in a consolidated amended complaint. Plaintiffs alleged that, among other things, Sprint’s ETF’s violated section 1671, subdivision (d) because they were “penalties,” which generated “substantial revenues and profits” and were intended “to prevent consumers from readily changing wireless telephone carriers.”8 The court granted Sprint *305leave to file a cross-complaint seeking monetary damages and equitable relief against class members for breach of contract in the event the RTF’s were found to be unenforceable penalties. The court denied Plaintiffs’ request to certify a cross-defendant consumer class on the basis that only setoff, and not affirmative relief, would be available if Sprint prevailed on its cross-complaint.9

By orders of December 10, 2007, and April 4, 2008, this case was severed and remanded for trial before Judge Bonnie Sabraw. In a March 17, 2008 pretrial order, the court considered which issues would be tried by the court and which by the jury. The court declined to bifurcate the case into separate court and jury trials, but identified the allocation of issues as follows: “First, the Court must decide whether RTFs are ‘rates’ under the Federal Communications Act (‘FCA’), 47 U.S.C. [§] 332(c)(3)(A). ...[][] Second, the Court must decide whether the RTFs are an alternative means of performance rather than a liquidated damage clause under the terms of the contracts at issue. . . . [f] Third, the Court must decide whether the RTFs of . . . [Sprint] are liquidated damage provisions under [section] 1671, and, if so, then whether they are lawful. ...[][] Finally, if the RTFs are unlawful, then a jury will determine the amount of damages under [section] 1671[, subdivision] (d), the CLRA, and the common count and the Court will determine the amount owed under the UCL and the claim for unjust enrichment.” Since the court anticipated significant overlap between the evidence relevant to both the court-tried issues and those the jury would be required to decide, it ruled that all issues would be presented in a single trial, and that the court and jury would then decide their respective issues at the conclusion of the evidence. By order dated April 17, 2008, the court denied Plaintiffs’ motion to try the issues of federal preemption, alternate performance, and invalidity of the RTF’s to the jury in an advisory capacity. Trial commenced on May 12, 2008.

Plaintiffs’ Evidence

Plaintiffs contended that the RTF’s were adopted and utilized by Sprint to stop erosion of its customer base by penalizing early termination of customer contracts, and as a revenue opportunity. The majority of Plaintiffs’ case was presented through the deposition testimony of Sprint employees, and through their expert Dr. Lee L. Selwyn.10

Testimony concerning Sprint’s initial decision to adopt a $150 RTF was presented by Plaintiffs through the video deposition of Bruce Pryor, Sprint’s *306vice-president of consumer marketing. In 1999, Sprint began to study the concept of term contracts with ETF’s as a means to reduce its “chum” rates,11 and tested use of ETF’s in selected markets. Sprint reported monthly wireless chum rates in 1998 of 3.3 percent, and in 1999 of 3.4 percent. Sprint adopted term contracts incorporating the $150 ETF nationwide in May 2000. Sprint reduced its chum rate to 2.8 percent in 2000.

The decision to implement ETF’s was made by Pryor and members of Sprint’s marketing team, including: Rob Vieyra, director of pricing; Chip Novick, vice-president of iharketing; Chuck Levine, chief marketing officer; and Andy Sukawaty, president of Sprint’s wireless division. Sprint had no surviving documentation relating to its decision to adopt ETF’s. Plaintiffs introduced contemporaneous Sprint internal documents referring to the ETF as a “$150 contract penalty fee,” and as a “Penalty or Contract Cancellation Fee.”

After Sprint’s August 2005 merger with Nextel, Sprint increased the amount of the early termination fee to $200. Sprint’s postmerger $200 ETF was based on Nextel’s premerger ETF. There was no evidence of any cost study made in connection with Nextel’s initial adoption of its $200 ETF (also in 2000), and Nextel did not prepare any written analysis of its decision to implement ETF’s.

It was undisputed that Sprint assessed ETF’s totaling $299,473,408 during the class period, and collected $73,775,975. Dr. Selwyn opined that, as a result of early contract terminations, Sprint avoided capital expenditures and variable costs, which were equal to about 98.6 percent of its monthly recurring charges. He calculated Sprint’s total lost profits from early terminations over the entire class period at $17,619,322.

Sprint’s Evidence

ETF’s are included in one-year and two-year term contracts, which offer heavily subsidized handsets and relatively low monthly charges, but are not included in month-to-month service plans. Sprint’s experts contended that an ETF is a part of the price the consumer pays for the “bundle” of the handset and cellular service, and is part of the quid pro quo for the rate reductions included in long-term plans. (See In re Ryder Communications, Inc. v. AT&T Corp. (2003) 18 F.C.C.R. 13603, ¶ 33.) Therefore, Sprint argued, any state law claim challenging use of ETF’s was preempted under federal law by the provisions of the federal Communications Act of 1934 (FCA; 47 U.S.C. § 151 et seq.), as amended in 47 United States Code section 332(c)(3)(A) (hereafter, section 332(c)(3)(A)).

*307To contest Plaintiffs’ claims that the ETF’s were unlawful liquidated damage provisions, and in support of its cross-complaint, Sprint sought to prove that its actual damages were substantially greater than the fees charged. Its trial evidence included information concerning Sprint’s costs, revenues, the frequency and timing of early terminations, and its efforts to collect ETF’s. Douglas M. Smith, Sprint’s chief technical operations officer, testified concerning Sprint’s network capacity. Wallace Souder, Jr., vice-president of pricing, testified as to Sprint’s costs and pricing practices, and Jay Michael Franklin, director of wireless service revenue, explained Sprint’s collections practices. Sprint presented three expert witnesses. Christian Dippon testified about the size of the Sprint ETF payer class, the timing of the contract terminations, and the revenues that Sprint contended that it lost as a result. Jeffrey Baliban gave evidence concerning costs that Sprint avoided as a result of the early terminations. Dr. William E. Taylor calculated the amount of damages Sprint claimed as a result of the early terminations. Sprint calculated that the payer class had 1,986,537 members; early terminations, on average, occurred with 13.25 months left on the term of the contract; and early terminations caused Sprint to lose $49.16 per month in monthly recurring charges. Baliban testified that costs avoided when a class member terminated early equaled about 18 percent of Sprint’s monthly recurring charges. Dr. Taylor opined that Sprint suffered damages of $987 million from early terminations, consisting of Sprint’s net revenue loss (monthly recurring charges lost minus costs avoided), less the amount of ETF’s actually collected.

The Jury Verdict

On June 12, 2008, the jury returned a verdict with special findings as follows: “1. What is the total dollar amount of early termination fees, that plaintiffs and the class members paid to Sprint? $73,775,975. [|] 2. Did plaintiffs and the class members breach their contracts with Sprint? Yes. [f] 3. State the total dollar amount of Sprint’s actual damages, if any, caused by early terminations of plaintiffs’ and class members’ contracts: $225,697,433.” The damages found by the jury were the exact amount of ETF’s charged to class members, but which were unpaid.

The Trial Court’s Statement of Decision

On December 4, 2008, after considering objections to its proposed statement of decision, the trial court issued its statement of decision. The court first reviewed the trial evidence presented and made its findings of fact. It initially accepted the jury’s determination of Sprint’s damages from early termination of consumer contracts in the amount of $225,697,433.

*308The court first held that section 332(c)(3)(A), denying states the authority to “regulate . . . the rates charged” by cellular telephone carriers, did not preempt Plaintiffs’ challenge to the ETF’s because the ETF’s were not “rates.”

The court found that the ETF in this case operated primarily as a liquidated damage clause. It also found Sprint’s ETF’s to be unenforceable penalties under section 1671, subdivision (d) because, although Sprint had established that it was impracticable to calculate the amount of actual damages from a breach at the inception of the contract, it had failed to meet its burden of establishing that it had made genuine and nonpretextual efforts to estimate a fair average compensation for the losses anticipated to be sustained (citing Hitz, v. First Interstate Bank (1995) 38 Cal.App.4th 274, 291 [44 Cal.Rptr.2d 890] (Hitz) [employing the reasonable endeavor test]).

The court determined that Sprint could not justify the ETF’s as a negotiated “alternative means of performance” under the contract, since they were invoked as liquidated damages upon a breach of the contract.

As a consequence of its determination that the ETF’s were unlawful under section 1671, subdivision (d), the court found that Plaintiffs had prevailed on their claims for violations of the CLRA (§ 1770, subd. (a)(14), (19)), UCL (Bus. & Prof. Code, § 17200 et seq.), unjust enrichment, and for money had and received. The court ordered restitution to the class in the amount of collected ETF’s ($73,775,975); enjoined Sprint from further efforts to collect ETF’s assessed during the class period; and ordered Sprint to advise third party assignees of uncollected claims of the court’s order. The court then, while questioning the validity of the jury’s verdict on damages, applied the setoff in favor of Sprint resulting from the jury’s verdict and determined that neither the class nor Sprint would be entitled to any monetary recovery. The setoff did not affect the injunctive relief granted.

Judgment was entered on December 24, 2008.

The Motion for New Trial

On December 15, 2008, Plaintiffs filed a notice of intention to move for a new trial on the jury’s verdict on questions 2 and 3, and the court’s calculation of the setoff. Plaintiffs contended that the court’s rulings and instructions had led the jury to presume the ETF’s were valid, to find breach of contract based on nonpayment of the ETF’s, and to award the amount of unpaid ETF’s instead of determining Sprint’s actual damages.

*309On January 27, 2009, the trial court granted Plaintiffs’ motion for new trial on the issue of actual damages (question 3) and on the setoff calculation. In granting the motion for new trial on the issue of Sprint’s actual damages (question 3), the court observed that it was “inconceivable that the jury considered the days of comprehensive and complex testimony by Dr. Selwyn, Mr. Baliban, and Dr. Taylor regarding Sprint’s lost revenue and avoidable costs and determined that Sprint’s actual total economic damages from all class members were exactly equal to the amount of unpaid ETF’s due from those class members who had not paid the ETF” and that the “finding that Sprint’s actual damages were $225,697,433 compels the conclusion that the jury did not follow the instructions to determine Sprint’s actual total economic damages.” It considered alternative methods that the jury could have used to reach its verdict and concluded that “[t]here is no way to read the jury’s answer to question #3 that yields a result that is both reasonable and lawful. The lawful readings are not reasonable and the reasonable readings are not lawful.” The grant of a new trial on damages necessarily required a new trial on the setoff to Sprint.

The court denied Plaintiffs’ motion for new trial on the issue of breach of contract by the class members (question 2), on the ground that there was substantial evidence in the record to support the jury’s implicit finding that the class members only incurred ETF’s as a consequence of breaching their contracts with Sprint.

The Appeals

On January 8, 2009, Sprint filed its notice of appeal from the December 24, 2008 judgment entered on the trial court’s statement of decision (appeal No. A124077). On February 5, 2009, Sprint filed its notice of appeal from the January 27, 2009 order granting in part Plaintiffs’ motion for new trial (appeal No. A124095). On February 17, 2009, Plaintiffs filed a timely notice of appeal from the December 24, 2008 judgment (appeal No. A125311), including the portion of the jury’s verdict finding that the class had breached its contracts with Sprint.

II. Discussion

A. Federal Preemption

“Congress has the power to preempt state law under the supremacy clause of the United States Constitution (art. VI, cl. 2). [Citation.] Congress’s intent to preempt may be expressly stated or implied where a federal law *310demonstrates an intent to ' “occupy the field” ’ or a state law conflicts with a federal law. [Citation.] A conflict exists where compliance with both state and federal law is impossible, or where a state law ‘ “stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress.” ’ [Citations.]” (Spielholz v. Superior Court (2001) 86 Cal.App.4th 1366, 1371 [104 Cal.Rptr.2d 197] (Spielholz).)

“To determine whether Congress intended to preempt state law with respect to a particular activity, we focus on the nature of the activity that the state seeks to control or regulate rather than on the method of regulation adopted. [Citation.] Preemption therefore applies not only to positive enactments by legislation or regulation but also to judicial acts that interfere or conflict with congressional intent. [Citation.] [|] Congress’s intent to preempt must be ‘clear and manifest’ to preempt state law in a field traditionally occupied by the states, such as the exercise of a state’s police powers.” (Spielholz, supra, 86 Cal.App.4th at pp. 1371-1372; see also Smith v. Wells Fargo Bank, N.A. (2005) 135 Cal.App.4th 1463, 1475 [38 Cal.Rptr.3d 653] (Smith).)

1. Standard of Review

The trial court considered and rejected the preemption claim. Finding the ETF to function primarily as a liquidated damage clause subject to California consumer protection statutes, the court applied a presumption against preemption. It found that Sprint had failed to meet its burden of demonstrating preemption, and that “[a] contractual agreement to replace a calculation of actual damages with liquidated damages does not transmute the calculation of contract damages (a traditional state function) into a wireless carrier’s ‘rate’ (a federal concern).”

In its opening brief, Sprint suggested that the preemption issue is entirely one of law, subject to de novo review. Plaintiffs assert that the court’s ruling was based primarily on findings of fact from the evidence presented at trial, and that our review is limited to examining the record for substantial evidence to support the court’s findings. In fact, they contend that Sprint has waived review of this issue by failing to address the evidence considered by the court, presenting only the evidence favorable to Sprint, and by failing to fully and fairly discuss the conflicting evidence. (See Huong Que, Inc. v. Luu (2007) 150 Cal.App.4th 400, 409 [58 Cal.Rptr.3d 527].) In its reply brief, Sprint acknowledges that the court’s ruling involved mixed questions of fact and law, but asserts that Sprint is challenging the legal standards employed by the trial court and the application of those standards to the trial court’s factual findings—not the findings themselves—and that such rulings are reviewable de novo.

*311Neither party is entirely correct. “The party claiming federal preemption bears the burden of establishing it. [Citation.]” (.Pacific Bell Wireless, LLC v. Public Utilities Com. (2006) 140 Cal.App.4th 718, 730 [44 Cal.Rptr.3d 733] (Pacific Bell Wireless).) “When the issues regarding federal preemption involve undisputed facts, it is a question of law whether a federal statute or regulation preempts a state law claim and, on appeal, we independently review a trial court’s determination on that issue of preemption. [Citations.]” (Smith, supra, 135 Cal.App.4th at p. 1476.) And insofar as the court resolved disputed issues of fact, its findings are reviewed under the substantial evidence standard, i.e., they will be sustained unless shown to lack substantial evidentiary support. (People ex rel. Gallo v. Acuna (1997) 14 Cal.4th 1090, 1136-1137 [60 Cal.Rptr.2d 277, 929 P.2d 596]; Howard S. Wright Construction Co. v. Superior Court (2003) 106 Cal.App.4th 314, 320 [130 Cal.Rptr.2d 641].)

2. Application of the FCA

The FCA “grants the Federal Communications Commission (FCC) broad authority over interstate and foreign communication by wire or radio, to secure and protect the public interest and to insure uniformity of regulation.” (Annot., Construction and Application of the Communications Act of 1934 and Telecommunications Act of 1996—United States Supreme Court Cases (2008) 32 A.L.R.Fed.2d 125.) In 1993 Congress amended the FCA (47 U.S.C. § 151 et seq., as amended by the Omnibus Budget Reconciliation Act of 1993, Pub.L. No. 103-66, § 6002 (Aug. 10, 1993) 107 Stat. 312, 387-397) to provide in relevant part that “no State or -local government shall have any authority to regulate the entry of or the rates charged by any commercial mobile service or any private mobile service, except that this paragraph shall not prohibit a State from regulating the other terms and conditions of commercial mobile services. . . .” (§ 332(c)(3)(A), italics added.)12

“By enacting section 332(c)(3)(A) in 1993, Congress ‘dramatically revise[d] the regulation of the wireless telecommunications industry, of which cellular telephone service is a part.’ [Citations.] ‘To foster the growth and development of mobile services [i.e., cellular and related mobile wireless communications] that, by their nature, operate without regard to state lines as an integral part of the national telecommunications infrastructure, new section 332(c)(3)(A) . . . preempts] state rate and entry regulation of all commercial mobile services,’ but permits state regulation of ‘other terms and conditions.’

*312(H.R.Rep. No. 103-111, 1st Sess., at p. 260, reprinted in 1993 U.S. Code Cong. & Admin. News, pp. 378, 587; 47 U.S.C. § 332.)” (Ball v. GTE Mobilnet of California (2000) 81 Cal.App.4th 529, 534 [96 Cal.Rptr.2d 801].) While only briefly discussing the meaning of “rate regulation,” the report of the House Budget Committee elaborated that “ ‘[b]y “terms and conditions,” the committee intends to include such matters as customer billing information and practices and billing disputes and other consumer protection matters; facilities siting issues (e.g., zoning); transfers of control; the bundling of services and equipment; and the requirement that carriers make capacity available on a wholesale basis or such other matters as fall within a state’s lawful authority. This list is intended to be illustrative only and not meant to preclude other matters generally understood to fall under “terms and conditions.” ’ ” (Cellico Partnership v. Hatch (8th Cir. 2005) 431 F.3d 1077, 1080, quoting H.R.Rep. No. 103-111, supra, 1993 U.S. Code Cong. & Admin. News, p. 588.)

Sprint contends that RTF’s are an integral part of its rate structure and that Plaintiffs’ claims are consequently expressly preempted by section 332(c)(3)(A). Plaintiffs argue that Sprint failed to meet its evidentiary burden of establishing that RTF’s are an element of its rates, and that RTF’s fall within the “other terms and conditions” provision of section 332(c)(3)(A), leaving them subject to state jurisdiction.

In Spielholz, cellular service provider AT&T contended that section 332(c)(3)(A) preempted state law claims of false advertising as to AT&T’s service coverage. (Spielholz, supra, 86 Cal.App.4th at p. 1369.) The class action plaintiffs in Spielholz sought injunctive and monetary relief, including damages and restitution. AT&T argued that for the court to award damages or restitution based on false advertising it must determine the value of the services provided and the reasonableness of the rates charged, and that such a determination would be expressly preempted as rate regulation within the meaning of section 332(c)(3)(A). (Spielholz, at pp. 1369-1370.)

The Court of Appeal first stayed resolution of appellate proceedings to allow the FCC (Federal Communications Commission) to consider a petition filed by Wireless Consumers Alliance, Inc., regarding the preemptive scope of section 332(c)(3)(A).13 (Spielholz, supra, 86 *313Cal.App.4th at p. 1370.) That petition before the FCC sought a declaratory ruling as to whether the FCA would preempt state courts from awarding monetary relief against cellular service providers: “(a) for violating state consumer protection laws prohibiting false advertising and other fraudulent business practices, or (b) in the context of contractual disputes and tort actions adjudicated under state contract and tort laws.” (In re Wireless Consumers Alliance, Inc. (2000) 15 F.C.C.R. 17021, f 1, fn. omitted (Wireless Consumers).) The FCC noted its prior declaratory ruling in In re Southwestern Bell Mobile Systems, Inc. (1999) 14 F.C.C.R. 19898, in which it found that the language and the legislative history of 47 United States Code section 332 did not support the preemption of state contract or consumer fraud laws relating to the disclosure of rates and rate practices. (Wireless Consumers, at f 14.) For the same reasons, it found that the language and legislative history of 47 United States Code section 332 did not support the view that state courts are, as a general matter, prevented by section 332(c)(3)(A) from awarding damages to customers of cellular service providers based on violations of state contract or consumer fraud laws. (Wireless Consumers, at f 14.) The FCC therefore determined that awarding monetary damages is not necessarily equivalent to rate regulation. (Id. at f 23.) “We agree with those commenters who contend that [47 United States Code sjection 332 was designed to promote the [commercial mobile radio service (CMRS)14] industry’s reliance on competitive markets in which private agreements and other contract principles can be enforced. It follows that, if CMRS providers are to conduct business in a competitive marketplace, and not in a regulated environment, then state contract and tort law claims should generally be enforceable in state courts.” (Wireless Consumers, at f 24, fn. omitted.) The FCC concluded: “A state court, by awarding damages to customers damaged by a CMRS provider’s breach of contract or fraud violation, would not per se be engaged in ratemaking prohibited by [47 United States Code s]ection 332(c)(3) . . . .” (Wireless Consumers, at f 38.) It further concluded that 47 United States Code section 332 “does not generally preempt state court award of monetary damages,” and thus there is “no inherent conflict between state common law or statutory remedies and the [FCA] . . . .” (Wireless Consumers, at 137.) The FCC cautioned, however, that while 47 United States Code section 332 does not generally preempt damage awards based on state contract or consumer protection laws, the question of whether a specific damage award or a specific grant of injunctive relief constitutes rate regulation prohibited by 47 United States Code section 332(c)(3) would depend on all facts and circumstances of the case. (Wireless Consumers, at f 39.)

Following the FCC’s decision, the Spielholz court vacated the stay and rejected the preemption claim. It observed that only rate regulation was *314directly prohibited by the FCA and held that express preemption did not apply even if a monetary damage award required determination of the value of AT&T’s service, since such a determination lacked a principal purpose and direct effect of controlling the provider’s rates and any effect on rates was merely incidental. (Spielholz, supra, 86 Cal.App.4th at pp. 1373, 1375-1376.) “In general, a claim that directly challenges a rate and seeks a remedy to limit or control the rate prospectively or retrospectively is an attempt to regulate rates and therefore is preempted under section 332(c)(3)(A); a claim that directly challenges some other activity, such as false advertising, and requires a determination of the value of services provided in order to award monetary relief is not rate regulation.” (Id. at pp. 1374-1375.)

The FCC has not yet ruled upon the question of whether ETF’s constitute “rates charged” under section 332(c)(3)(A), such that state law claims are preempted.15 Only one California appellate case has addressed ETF’s at all, albeit in a slightly different context. In Pacific Bell Wireless, California’s Public Utilities Commission (PUC) imposed fines and restitution orders against Pacific Bell Wireless, LLC, doing business as Cingular Wireless (Cingular) for, among other things, what the PUC found to be the unjust and unreasonable practice of charging its customers an ETF without permitting any type of grace period. (Pacific Bell Wireless, supra, 140 Cal.App.4th at p. 723.) The court rejected Cingular’s claim that the PUC decision was preempted by section 332(c)(3)(A). “The [PUC’s] decisions do not directly challenge Cingular’s rates, nor do they require Cingular to make any specific changes to its infrastructure. As in Spielholz . . . , the [PUC’s] challenge to the ETF and to Cingular’s policy of permitting no grace period, *315combined with the misrepresentations regarding service, is not a preempted regulation of rates or of market entry. The principal purpose and direct effect of the penalties' imposed by the [PUC] are to prevent misrepresentations by Cingular and to compensate the wireless customers who paid ETF’s. The effect of these penalties on Cingular’s rates is incidental, and the [PUC’s] decisions are therefore not preempted by . . . section 332(c)(3)(A).” (Pacific Bell Wireless, at p. 734.) While Plaintiffs argue that the court in the Pacific Bell Wireless decision “definitively determined” that Cingular’s ETF was not a “ ‘rate charged’ ” (Pacific Bell Wireless, at p. 732) for service under section 332(c)(3)(A), the precise issue before the court, as Sprint points out, was whether the PUC’s orders were preempted by section 332(c)(3)(A), not whether the ETF’s were rates within the meaning of the statute. The PUC, in its decision, focused upon the conditions under which Cingular imposed the ETF. It made “no findings on whether imposition of an ETF [was] unreasonable per se,” or “what amount, if any, [would] constitute[] a reasonable or unreasonable ETF.” (Pacific Bell Wireless LLC doing business as Cingular Wireless (Cal.P.U.C. Sept. 23, 2004) Dec. No. 04-09-062 [2004 Cal.P.U.C. Lexis 453, pp. *77-*78].) The PUC held that Cingular’s legal culpability stemmed from imposing a no retum/no refund ETF from day one of the contract period without providing any trial period. (Id. at p. *125.) Neither the PUC nor the Court of Appeal determined whether the ETF itself was a rate charged subject to federal preemption.

Nor have we found reported appellate authority from any other jurisdiction on this point. Contrary to Plaintiffs’ assertion, there is no “overwhelming weight of authority” that state law challenges to ETF’s are not preempted rate regulation. Trial courts considering the question have reached conflicting conclusions.16 (See, e.g., Phillips v. AT&T Wireless (S.D. Iowa, July 29, 2004, No. 4:04-cv-40240) 2004 U.S.Dist. Lexis 14544, pp. *35-*37; Iowa v. United States Cellular Corp. (S.D. Iowa, Aug. 7, 2000, No. 4-00-CV-90197) 2000 U.S.Dist. Lexis 21656, p. *20; Cedar Rapids Cellular Telephone LP v. Miller (N.D. Iowa, Sept. 15, 2000, No. C00-58 MJM) 2000 U.SDist. Lexis 22624, pp. *20-*21 [holding that an ETF is not a rate]; Esquivel v. Southwestern Bell Mobile Systems, Inc. (S.D.Tex. 1996) 920 F.Supp. 713, 715 [“liquidated damage provision here is a ‘term and condition’ of the agreement rather than a rate”]; cf. Redfern v. AT&T Wireless Services, Inc. (S.D.Ill., June 16, 2003, Civ. No. 03-206-GPM) 2003 U.S.Dist. Lexis 25745, pp. *2-*3 [challenge to ETF preempted since ETF’s are “ ‘an integral part of the rates charged by [the defendant] for its services’ under its wireless service agreements”]; Chandler v. AT&T Wireless Services, Inc. (S.D.Ill., July 21, 2004, *316Civ. No. 04-180-GPM) 2004 U.S.Dist. Lexis 14884, p. *4 [ETF charges “directly connected to the rates charged for mobile services”].) Other courts have elected to stay state law based ETF challenges awaiting FCC guidance on the issue. (See Waudby, supra, 2007 U.S.Dist. Lexis 38581 at pp. *18—*19; Gentry v. Cellico P’ship (C.D.Cal., Mar. 22, 2006, No. CV 05-7888 GAF (VBKx)) 2006 U.S.Dist. Lexis 97876.)

a. Factual Issues

As we have discussed, Sprint’s position is that this court should review the trial court’s rejection of the preemption defense de novo.17 It is clear that the trial court considered the question of federal preemption here as a mixed issue of law and fact. So do we.

We first observe that while the FCC has still to address the ETF question, it has suggested that the issue of rate preemption under section 332(c)(3)(A) is, at least to some degree, fact intensive. (Wireless Consumers, supra, 15 F.C.C.R. 17021, at f 39.) While concluding that there is “no inherent conflict between state common law or statutory remedies and the [FCA],” the FCC nevertheless cautioned that the question of whether a specific damage award or a specific grant of injunctive relief based on state contract or consumer protection laws constitutes rate regulation prohibited by section 332(c)(3)(A) would depend on all facts and circumstances of the case. (Wireless Consumers, at fl 37, 39.)

The trial court took this approach. By order of January 20, 2004, the original coordination judge overruled a demurrer to the ETF claims based on an argument of federal preemption by section 332(c)(3)(A), finding that the factual allegations of the pleadings were “unsettled,” and that evaluation of whether ETF’s were rates “will require a decision based on consistent pleadings or an evidentiary record on summary judgment or at trial. [Citation.]” In its March 17, 2008 pretrial order, the trial judge denied Plaintiffs’ *317motion for judgment on the pleadings on the federal preemption affirmative defense, stating that “[t]he court is not persuaded that it can resolve the preemption issue as a pleading issue any more now than it could in 2004.”18 The court said that it would decide at trial if the ETF’s were rates under section 332(c)(3)(A).

Sprint contends that undisputed evidence at trial showed that its ETF’s are an element of the prices it charged, and that the ETF’s are an “integral part of Sprint’s rate structure.” Sprint points to testimony of one of its experts (Taylor) that ETF’s are “potentially one of the prices you pay” under Sprint’s cell phone plans. According to Taylor, cell phone service is a package with separate prices for its individual components, and just as customers are charged under some contracts if they decide to “roam” outside a geographic area, “if you decide to leave early, there may be an early termination price.” Sprint contends that Plaintiffs’ expert, Dr. Selwyn, agreed, conceding on cross-examination that ETF’s “are a form of payment for the handsets and cellular service that Sprint provides with its long-term contracts,” with customers paying for the handset “in three ways: up-front charges for the contract, recurring monthly service charges, and what Selwyn characterized as the ‘liquidated damages payment in the event of breach of the installment contract.’ ” Sprint says that the evidence shows that ETF’s are an integral part of its rate structure because they are included in one-year and two-year term contracts, which offer heavily subsidized handsets and relatively low monthly charges, but not in its month-to-month service plans. Therefore, ETF’s are, they claim, part of the “quid pro quo for the rate reductions included in long-term plans.” (In re Ryder Communications, Inc. v. AT&T Corp., supra, 18 F.C.C.R. 13603, at ¶ 33.)

As Plaintiffs correctly observe, however, Sprint ignores the evidence considered by the trial court, and set forth in its statement of decision, as to the “factual nature of the ETF.” The trial judge discussed the evidence concerning the “true nature” of Sprint’s ETF, “which the Court discerns by looking at its objective effect on commerce generally and its effect on the majority of Sprint’s customers specifically,” and found that “Sprint’s ETF operated primarily as a liquidated damage clause.” The court cited the testimony of Bruce Pryor, who testified that Sprint’s goal in adopting the ETF was to control chum and was implemented, as was Nextel’s similar charge, “primarily as a means to prevent customers from leaving.” It noted that Sprint “did no analysis that considered the lost revenue from contracts, the avoidable costs, and Sprint’s expected lost profits from contract terminations.” The ETF’s were set “from a competitive standpoint,” and “Sprint’s early evaluations of the ETF assumed that Sprint would not collect any money from the *318ETFs.” The trial court also found that several different versions of Sprint’s form subscriber agreements all referred to the ETF as “liquidated damages.” Sprint did not challenge the court’s findings in its opening brief, and expressly states in its reply brief that it “is challenging the legal standards employed by the trial court and the application of those standards to the trial court’s factual findings, not the findings themselves.” We find, in any event, that the court’s findings are supported by substantial evidence.

b. Presumption Against Preemption

The interpretation and application of the federal law at issue “is further informed by a strong presumption against preemption. [Citations.] ‘[B]ecause the States are independent sovereigns in our federal system, we have long presumed that Congress does not cavalierly pre-empt state-law causes of action. In all pre-emption cases, and particularly in those in which Congress has “legislated ... in a field which the States have traditionally occupied,” [citation], we “start with the assumption that the historic police powers of the States were not to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress.” [Citations.]’ [Citations.]” (Farm Raised Salmon Cases, supra, 42 Cal.4th at p. 1088; see Smith, supra, 135 Cal.App.4th at p. 1475.) “We apply this presumption to the existence as well as the scope of preemption. [Citation.]” (Farm Raised Salmon Cases, at p. 1088.) “Considering the general presumption against preemption, we narrowly construe the precise language of the federal law or regulation to determine whether a particular state law claim is preempted. [Citations.]” (Smith, at p. 1476.)

Sprint argues that the presumption against preemption is not triggered when the state regulates in an area, like communications law, where there has been a history of significant federal presence. This again is an “implied preemption” argument not applicable here. In Spielholz, the court rejected a claim of implied preemption, finding that the FCA does not evidence an intent to “occupy the field” by precluding other remedies than petition to the FCC to address a service provider’s classifications, practices and regulations, and that a potential federal remedy is not necessarily inconsistent with state remedies, as acknowledged by the saving clause within the statute. (Spielholz, supra, 86 Cal.App.4th at p. 1376.) “Moreover, the availability of state law remedies is consistent with the 1993 amendments’ objective to achieve maximum benefits for consumers and providers through reliance on the competitive marketplace, in which state law duties and remedies ordinarily are enforceable. [Citation.]” (Id. at pp. 1376-1377, citing Wireless Consumers, supra, 15 F.C.C.R. 17021, at ff 22, 24.) Furthermore, even without a presumption *319against preemption, “because preemption of state laws by federal law or regulation generally is not favored, the party claiming federal preemption . . . has the burden to show specific state law claims are preempted. [Citations.]” (Smith, supra, 135 Cal.App.4th at pp. 1475-1476, fn. omitted.)

c. FCA Does Not Preempt Application of California Law to Sprint’s ETF’s

The trial court found that Sprint failed to meet its burden of establishing preemption. We agree.

There is a distinction between rates that are filed with an agency and subject to public and regulatory review, and prices that are determined and published by the carrier in a competitive marketplace.19 (Wireless Consumers, supra, 15 F.C.C.R. 17021, at ][ 20.) Sprint argues that it subsidizes its charges to customers for equipment and sets its charges for service on the assumption that it will recover its costs over the term of its fixed contracts, and that the ETF’s compensate it for its losses when a customer fails to fulfill the full contract term. But this argument actually confirms that Sprint’s rates for equipment and services are established at a level to provide its full projected return on investment over the contract term, assuming that the customer fulfills his or her obligations. Only if a customer failed to do so would Sprint suffer any loss of anticipated revenue, with the amount of that loss dependent upon when the customer default occurred.

But the ETF’s were not prorated, so that the customer would pay the same amount whether the termination occurred during the first month or the last *320month of the contract term. More significantly, the court found that Sprint, in implementing ETF’s, “did no analysis that considered the lost revenue from contracts, the avoidable costs, and Sprint’s expected lost profits from contract terminations,” and that “Sprint’s early evaluations of the ETF assumed that Sprint would not collect any money from the ETFs.”20 The ETF’s were not based on the amount of any actual or estimated loss, but “from a competitive standpoint.” Sprint’s purpose in adopting the ETF was to control chum and was implemented “primarily as a means to prevent customers from leaving.” In other words, the ETF’s were intended not to be a collectible element of Sprint’s rates, but rather to serve as a deterrent—either coercing customer compliance with the contract rate structure or penalizing noncompliance. But as we discuss post, Sprint mns afoul of California consumer protections in doing so. Simply labeling an ETF as a rate because it is charged to certain customers does not make it one.

It is certainly possible that elimination of ETF’s may indirectly affect Sprint’s rates to the extent that Sprint incurs costs in pursuing alternative remedies for contractual breach or that it would reserve for losses attributable to a potentially higher level of customer defaults. Sprint would presumably factor actual or projected lost revenue into its rate structure. We agree with the Spielholz court, however, that “[rjate regulation, or to ‘regulate ... the rates charged’ in the words of section 332(c)(3)(A), refers ... to an action whose principal purpose and direct effect are to control prices. ... [f] ... [1] A judicial act constitutes rate regulation only if its principal purpose and direct effect are to control rates.” (Spielholz, supra, 86 Cal.App.4th at pp. 1373-1374.) In Ball v. GTE Mobilnet of California, the court, while finding federal preemption of state law claims contesting “rounding up” of per minute charges for wireless calls, distinguished cases involving “billing practices” having “only a tangential relationship to the actual rates for service paid by cellular customers” (including Esquivel v. Southwestern Bell Mobile Systems, Inc., supra, 920 F.Supp. 713 [finding a charge for early termination of cellular service to be a “term and condition” of service, not a rate, and therefore subject to state regulation]). (Ball v. GTE Mobilnet of California, supra, 81 Cal.App.4th at pp. 539, 541.)

Although not directly on point, we also find the reasoning in Pacific Bell Wireless, in a closely related context, persuasive. Upholding imposition of penalties and a restitution order issued by the PUC against Cingular for, *321among other things, charging ETF’s without any grace period for cancellation, our colleagues in the Fourth District Court of Appeal concluded that the PUC’s decisions were not preempted by the FCA because they did not “directly challenge Cingular’s rates, nor do they require Cingular to make any specific changes to its infrastructure.” (Pacific Bell Wireless, supra, 140 Cal.App.4th at p. 734.) “The principal purpose and direct effect of the penalties imposed by the [PUC] are to prevent misrepresentations by Cingular and to compensate the wireless customers who paid ETF’s. The effect of these penalties on Cingular’s rates is incidental, and the [PUC’s] decisions are therefore not preempted by [the FCA].” (Ibid.)

The trial court correctly held that “[a] contractual agreement to replace a calculation of actual damages with liquidated damages does not transmute the calculation of contract damages (a traditional state function) into a wireless carrier’s ‘rate’ (a federal concern).” As the FCC has found, section 332(c)(3)(A) “was designed to promote the [wireless cellular] industry’s reliance on competitive markets in which private agreements and other contract principles can be enforced. It follows that, if [cellular] providers are to conduct business in a competitive marketplace, and not in a regulated environment, then state contract and tort law claims should generally be enforceable in state courts.” (Wireless Consumers, supra, 15 F.C.C.R. 17021, at ][ 24, fn. omitted.)

Invalidation of the ETF’s under California’s consumer protection laws will have only an indirect and incidental effect on Sprint’s rates and, therefore, is not preempted by section 332(c)(3)(A).

B. Section 167121

The trial court found that an ETF operated primarily as a liquidated damage clause. Because Sprint failed to prove that, in adopting ETF’s, it made any effort “to determine what losses it would sustain from breach by the early termination of its contracts” or “to estimate a fair average compensation for such losses,” it failed to satisfy the reasonable endeavor test and the ETF’s were consequently unlawful penalties under section 1671, subdivision (d).

A provision in a consumer contract “liquidating damages for the breach of the contract is void except that the parties to such a contract may agree therein upon an amount which shall be presumed to be the amount of damage sustained by a breach thereof, when, from the nature of the case, it would be impracticable or extremely difficult to fix the actual damage.” *322(§ 1671, subd. (d); see id., subd. (c)(1).) Because liquidated damage clauses in consumer contracts are presumed void, the burden is on the proponent of the clause to rebut that presumption. (Garrett v. Coast & Southern Fed. Sav. & Loan Assn. (1973) 9 Cal.3d 731, 738 [108 Cal.Rptr. 845, 511 P.2d 1197] (Garrett).)

Decisions interpreting this statute have created a two-part test for determining whether a liquidated damages provision is valid: (1) fixing the amount of actual damages must be impracticable or extremely difficult, and (2) the amount selected must represent a reasonable endeavor to estimate fair compensation for the loss sustained. (Utility Consumers’ Action Network, Inc. v. AT&T Broadband of Southern Cal., Inc. (2006) 135 Cal.App.4th 1023, 1029 [37 Cal.Rptr.3d 827] (Utility Consumers).) “Absent either of these elements, a liquidated damages provision is void . . . .” (Hitz., supra, 38 Cal.App.4th at p. 288, italics added.) A liquidated damages provision need not, however, be expressly negotiated by both parties to a form contract in order to be valid. (Utility Consumers, at p. 1035.)

Impracticability may be established by showing “that the measure of actual damages would be a comparatively small amount and that it would be economically impracticable in each instance of a default to require a [seller] to prove to the satisfaction of the [consumer] the actual damages by accounting procedures.” (Garrett, supra, 9 Cal.3d at p. 742.) The trial court found that, although Sprint could readily calculate its lost monthly revenue per customer in the event of a default, it would have been impracticable to determine Sprint’s avoidable costs, and therefore impracticable to determine actual damages at the inception of the contracts. Plaintiffs do not challenge this finding.

“Determining whether a reasonable endeavor was made depends upon both (1) the motivation and purpose in imposing the charges, and (2) their effect.” (Utility Consumers, supra, 135 Cal.App.4th at p. 1029.) “[T]he focus is not ... on whether liquidated damages are disproportionate to the loss from breach, but on whether they were intended to exceed loss substantially—a result of which is to generate a profit.” (Hitz, supra, 38 Cal.App.4th at p. 289.) A liquidated damages provision that “bears no reasonable relationship to the range of actual damages that the parties could have anticipated would flow from a breach” is an unlawful penalty that compels a forfeiture upon a breach of contract. (Ridgley v. Topa Thrift & Loan Assn. (1998) 17 Cal.4th 970, 977 [73 Cal.Rptr.2d 378, 953 P.2d 484].) Such penalties are “ ‘ineffective, and the wronged party can collect only the actual damages sustained.’ [Citations.]” (Id. at p. 977.)

In order to establish the reasonable endeavor required, evidence must exist that the party seeking to impose liquidated damages “ ‘actually engaged in *323some form of analysis to determine what losses it would sustain from [a] breach, and that it made a genuine and non-pretextual effort to estimate a fan-average compensation for the losses to be sustained.’ ” (Hitz, supra, 38 Cal.App.4th at p. 291.) The trial court made a finding that “when Sprint implemented the ETF in 2000, and increased it in 2005, it made no endeavor—reasonable or otherwise—to determine what losses it would sustain from breach or to estimate a fair average compensation for such losses.” Sprint “did no analysis that considered the lost revenue from contracts, the avoidable costs, and Sprint’s expected lost profits from contract terminations.” The ETF amounts were set not based on the basis of any actual or estimated loss, but “from a competitive standpoint.” Sprint’s purpose in adopting the ETF was to control chum and was implemented “primarily as a means to prevent customers from leaving.” After adoption of ETF’s, Sprint succeeded in reducing its chum rate to 2.8 percent in 2000.

As discussed above, the court’s findings are supported by substantial evidence. The testimony of Bruce Pryor, Sprint’s vice-president of consumer marketing, was that Sprint began to study the concept of term contracts with ETF’s in 1999 as a means to reduce its chum rates. The decision to implement ETF’s was made by members of Sprint’s marketing team. Contemporaneous Sprint internal documents referred to the ETF as a “$150 contract penalty fee,” and as a “Penalty or Contract Cancellation Fee.” Sprint’s postmerger $200 ETF was based on Nextel’s premerger ETF. The trial court found “no evidence at trial that Nextel did any analysis that considered the lost revenue from contracts, the avoidable costs, or Nextel’s expected lost profits from contract terminations.”

Sprint counters that “undisputed evidence” at trial showed that the ETF’s were not intended to exceed losses, and that Sprint officials were “aware that their ETFs would recover only a fraction of the revenue lost as a result of early terminations.” Sprint asserts that any charge that “does not overstate actual damages cannot be a penalty.” Sprint cites testimony that early terminations occurred on average with 13.25 months left on a contract, depriving Sprint of average revenues of $49.16 per month in monthly recurring charges per customer and that it lost over $650 in revenue for each early termination. Sprint points out that the ETF’s do not even cover their costs of adding new customers, which averaged $388 dollars during the class period. It contends that the evidence showed that it “well understood that due to competitive forces, the ETF could not be set anywhere near a level that would compensate it for a customer’s breach through early termination.”

*324Sprint’s expert calculated that Sprint’s actual damages from early terminations by the class members was $987 million.22 Sprint contends that the ETF’s on average reduced each class member’s net obligation under their contract by more than $450, and that a subscriber with a $150 ETF would have reduced his or her net payments to Sprint by voluntarily paying the ETF with at least four months remaining on the contract, and a subscriber with a $200 ETF could have done the same by voluntarily paying an ETF with at least five months left. As we discuss further post, the jury found that Sprint’s actual damages from early terminations were exactly equal to the amount of its uncollected ETF’s ($225,697,433).

Sprint asserts that the trial court erroneously failed to consider the effect of Sprint’s ETF’s and in requiring a “formal study of estimated damages” to satisfy the motive-and-puipose prong of the reasonable endeavor test. Sprint argues that because the ETF’s were not “intended to exceed loss substantially,” they do not and cannot violate the motive-and-purpose aspect of the reasonable endeavor test. (Hitz, supra, 38 Cal.App.4th at p. 289.)

We note first that, as to Sprint’s motive and purpose, whatever information as to costs and revenues Sprint may have been “aware” of, it cites to no evidence in the trial record that any of this information was part of the calculus in deciding to impose ETF’s, or in determining the amount of an ETF. Further, we do not second-guess the trial court’s factual determinations as to Sprint’s motivation and purpose. (Hitz, supra, 38 Cal.App.4th at p. 290.)

While Sprint contends that it satisfied both prongs of the reasonable endeavor test in adopting ETF’s, the real focus of its argument is on the effect of the ETF’s. The thrust of that argument is that so long as the ETF amount is shown in practice to be less than Sprint’s actual damages, the effect is not to generate a profit, whatever Sprint’s motive and purpose, and nothing more is required to meet the test.

Sprint further contends that this court, in our discussion of section 1671, subdivision (d) in our prior unpublished decision in In re Cellphone Termination Fee Cases, supra, A115457, so held.23 Sprint seeks to cull a holding from our prior decision which we did not make. In that decision we were required to consider whether the court abused its discretion in denying certification of a current subscriber class in this case on the basis of potential *325intraclass conflicts.24 While we reviewed the parties’ respective positions on liquidated damages, we held that consideration of the merits of Plaintiffs’ claims (that ETF’s were invalid under § 1671, subd. (d)) was improper on a motion for class certification. We did not hold, as Sprint contends, that “Sprint’s ETF’s would ‘satisfy [] the reasonable endeavor requirement . . .’if they approximated actual damages,” nor did we hold that if early termination fees proved to be “ ‘reasonable estimates of Defendants’ damages, [that] would ‘mean[] they were not unlawful penalties.’ ” “An appellate decision is not authority for everything said in the court’s opinion but only ‘for the points actually involved and actually decided.’ [Citations.]” (Santisas v. Goodin (1998) 17 Cal.4th 599, 620 [71 Cal.Rptr.2d 830, 951 P.2d 399].)

Sprint likewise places undue reliance on our Supreme Court’s decision in Better Food Mkts. v. Amen Dist. Teleg. Co. (1953) 40 Cal.2d 179 [253 P.2d 10] (Better Food), contending that Better Food imposes an entirely objective effect test. The plaintiff in Better Food was a grocery company that contracted with the defendant to install a burglar alarm system, then monitor that system, and notify the police if the alarm were triggered. It sued for damages when the defendant failed to do so and the store suffered a large loss. (Id. at p. 182.) The Supreme Court reversed a directed verdict for the defendant, but also held that any recovery for the plaintiff would be limited to $50, the amount set forth in the form contract’s liquidated damages provision. (Id. at p. 188.) The court reiterated the rule that the amount of liquidated damages “must represent the result of a reasonable endeavor by the parties to estimate a fair average compensation for any loss that may be sustained.” (Id. at p. 187.) It also held that even though the liquidated damages provision was found in a form contract, and even though the defendant did not investigate the plaintiff’s manner of doing business or the character and value of its stock, “the parties agreed to the liquidation provisions, and there is no evidence that they were not fully aware of circumstances making it desirable that liquidated damages be provided for.” (Ibid.) The court in Better Food, however, dealt with a commercial contract and focused on the impracticability of fixing actual damages and the parties’ agreement to the liquidated amount at the time of contracting. (Id. at p. 187.) It also did not articulate any rationale or underlying policy behind its rulings. (Utility Consumers, supra, 135 Cal.App.4th at p. 1038.)

The Utility Consumers case lends some support to Sprint’s position. That court, in rejecting a claim that actual mutual negotiation was required to validate liquidated damages in a consumer contract (in that instance, late fees), made a detailed analysis of the cases underpinning and articulating the reasonable endeavor test. (Utility Consumers, supra, 135 Cal.App.4th at *326pp. 1029-1039.) The court concluded that “the reasonable endeavor test looks primarily to the intent of the parties, as determined by the purposes behind a liquidated damages clause and the relationship between the amount of liquidated damages and a fair estimate of the actual damages from a breach of the contract. [Citations.]” (Id. at p. 1038.) The Utility Consumers court was critical of what it viewed as the interpretation of Garrett, supra, 9 Cal.3d 731 in Hitz, supra, 38 Cal.App.4th at page 289, “as focusing solely on the intent behind a liquidated damages provision, not on whether the amount selected was disproportionate to the loss from the breach.” (Utility Consumers, at p. 1038, fn. 9, italics added.) The court observed that “ ‘the characteristic feature of a penalty is its lack of any proportionate relation to the damage which may actually stem from the breach of a contract [citations].’ [Citation.]” (Id. at p. 1031, quoting McCarthy v. Tally (1956) 46 Cal.2d 577, 584, fn. 4 [297 P.2d 981].) In Utility Consumers, however, the plaintiff stipulated for purposes of the summary judgment motion that the defendant had performed an analysis to determine its actual late payment costs, and that, according to the analysis, those costs were greater than the late fee imposed, and conceded for purposes of the motion that the late fee charged was reasonable. (Utility Consumers, at p. 1026.) Further, the Utility Consumers court expressly did not reach the issue of whether the intent behind a liquidated damages provision could be fatal to a reasonable endeavor analysis, whether or not the amount of liquidated damages was disproportionate to the loss from the breach. (Id. at p. 1038, fn. 9.)

Here, as the trial court found, the evidence fails to establish any endeavor, reasonable or otherwise, to even approximate Sprint’s actual damages flowing from breach of the term contracts by consumers, and instead reflects a marketing decision made with an entirely deterrent purpose and focus. We believe that Plaintiffs are correct that the reasonable endeavor test, to have any meaning, must necessarily focus on those circumstances actually considered in evaluating a liquidated damage provision, not post hoc rationalization.

In Hitz, the plaintiff class challenged late and overlimit fees on credit card balances. (Hitz, supra, 38 Cal.App.4th at pp. 277-278.) There was trial evidence that the defendant bank was looking for new sources of revenue, and the bank’s witnesses admitted when the bank made the decision to impose the fees, it had conducted no study or analysis of the costs resulting from late and overlimit activity. (Id. at p. 289.) The bank nevertheless relied' upon testimony from the responsible bank officer that he “ ‘had a good understanding of our costs from the information that I got on a regular basis,’ and on his assertion that he never ‘saw a situation’ where [the bank] made a profit from late and overlimit fees.” (Id. at p. 290.) The court observed that the “pivotal factor” was the bank’s motivation and purpose, not whether an officer personally had a “ ‘good understanding’ of costs.” (Ibid.) In rejecting an argument similar to that Sprint makes here, the Hitz court found the bank’s *327reliance on cost studies from later years to be not pertinent to its motivation and purpose when it decided the amounts of its late and overlimit fees, and “irrelevant to the reasonable endeavor issue.” (Id. at pp. 291-292.) This is “because the validity of liquidated damages ‘is determined by circumstances existing when the fee provisions are inserted into the contract, and not by subsequent events ....’” (Id. at p. 291.) Hitz held that “The ‘amount’ of liquidated damages ‘must represent the result of a reasonable endeavor’ to estimate fair compensation. [Citation.] For the amounts of the challenged fees to have been such a result, the required reasonable endeavor logically must have preceded the setting of those amounts.” (Ibid., quoting Garrett, supra, 9 Cal.3d at p. 739.)

We see no necessary conflict between Hitz and Utility Consumers. We think that Utility Consumers correctly acknowledged that “whether a reasonable endeavor was made depends upon both (1) the motivation and purpose in imposing the charges, and (2) their effect.” (Utility Consumers, supra, 135 Cal.App.4th at p. 1029, italics added.) Failure to meet its burden of proof on either element is fatal to Sprint’s position.

Sprint again insists that the evidence shows that the ETF’s benefit consumers by allowing Sprint to offer reduced monthly fees and subsidized handsets, as well as generally imposing charges less than, or at least equal to, Sprint’s actual damages. Sprint contends that applying the reasonable endeavor test in these circumstances will expose consumers to liability for higher actual damages and would be inconsistent with the underlying rationale of the rule. But focusing solely on hindsight justifications would render the reasonable endeavor requirement meaningless if no effort at foresight were required, and arbitrarily selected charges could be routinely imposed in consumer contracts, subject only to the ability of a company to muster a credible defense if challenged in litigation. If no attempt to make a reasonable assessment of potential loss is required at the outset, and to make the amount of the liquidated damages consonant with that assessment, one of the important functions that liquidated damages serve, removing the uncertainty factor from determining damages from a breach of contract and reducing litigation, would be lost. (See Utility Consumers, supra, 135 Cal.App.4th at p. 1038.)

Sprint was required to show that it actually engaged in some form of analysis to determine what losses it would sustain from breach,25 and that it *328made a genuine effort to estimate a fair average compensation for the losses to be sustained. Sprint may be correct that in retrospect its ETF’s were reasonable in amount in light of its actual loss, and that they may actually have been beneficial in practice to at least some of its customers. However, institutional intuition is not a substitute for analytical evaluation, and retrospective rationalization does not excuse the objective assessment required at the inception of the contract.

C. Alternative Performance

Sprint also sought to defend use of ETF’s as “alternative performance,” permitting subscribers to terminate contracts before the end of the agreement by paying a fee. “[T]o constitute a liquidated damage clause the conduct triggering the payment must in some manner breach the contract.” (Morris v. Redwood Empire Bancorp (2005) 128 Cal.App.4th 1305, 1315 [27 Cal.Rptr.3d 797] (Morris).) A contractual provision that merely provides an option of alternative performance of an obligation does not impose damages and is not subject to section 1671 limitations. (See Garrett, supra, 9 Cal.3d at p. 735.)

In evaluating the legality of a provision, a court must first determine its true function and operation. (Garrett, supra, 9 Cal.3d at p. 735; Morris, supra, 128 Cal.App.4th at p. 1315.) “[W]hen it is manifest that a contract expressed to be performed in the alternative is in fact a contract contemplating but a single, definite performance with an additional charge contingent on the breach of that performance, the provision cannot escape examination in light of pertinent rules relative to the liquidation of damages. [Citations.]” (Garrett, at p. 738.) To hold otherwise “would be to condone a result which, although directly prohibited by the Legislature, may nevertheless be indirectly accomplished through the imagination of inventive minds.” (Id. at p. 737.)

As Plaintiffs note, Sprint itself, in several different versions of Sprint’s form subscriber agreements referred to the ETF as a “liquidated damages” provision. The trial court “categorize[d] and analyze[d] the ETF by looking at its true nature, which the Court discem[ed] by looking at its objective effect on commerce generally and its effect on the majority of Sprint’s customers specifically. [Citations.]” The court found that the ETF provisions “did not give customers a rational choice of paying the ETF or completing the contract,” because the language of the ETF provision permitted Sprint to impose the fee on customers involuntarily. The court noted that “[o]f those customers who were charged an ETF, 80% were terminated by Sprint and experienced the ETF as the imposition of liquidated damages . . . .” As the *329trial court stated, “If this case concerned a Sprint clause that stated customers could terminate term contracts early by paying a fee, then that fee might well be an alternative means of performance.” Instead, “Sprint declared contracts breached, terminated service, and imposed ETFs as liquidated damages resulting from the asserted breaches.”

Sprint argues that the trial court erred in judging the economic function of the ETF and choice it provided customers after the contract had either been performed or breached, and that it should instead have judged the choice the ETF provided customers at the outset of the contract. (Blank v. Borden (1974) 11 Cal.3d 963, 971 [115 Cal.Rptr. 31, 524 P.2d 127] [arrangement viewed from the time of making the contract].) But, as Plaintiffs respond, the service agreements provided from the inception of the contract that an ETF could be triggered involuntarily by Sprint, confirming that at the time of contracting the provision was not understood or intended as providing only for a “rational choice” of the customer.

Sprint cites a trial level decision of the United States district court in Hutchison v. AT&T Internet Services, Inc. (C.D.Cal., May 5, 2009, No. CV073674 SVW (JCx)) 2009 U.S.Dist. Lexis 53937 (Hutchison). There the trial judge granted summary judgment to an Internet service provider on a claim that its ETF violated section 1671. The court found that the fee there provided a realistic and rational choice of alternative performance to the subscriber. (2009 U.S.Dist. Lexis 53937 at pp. *12-*14.) While the federal trial court determination is not binding on this court in any event, it is self-evident that in contrast we deal here with contrary factual findings made after trial on a full evidentiary record.26

The court found that “Sprint has not met its burden of establishing that the predominant effect of the ETF provisions was to provide consumers with an alternate means of performing their contracts.” Substantial evidence supports the findings.

D., E.*

*330III. Disposition

The judgment of the trial court is affirmed. The matter is remanded for retrial on the issue of Sprint’s damages, and the calculation of any offset to which Sprint may be entitled. Neither party shall recover costs on this appeal.

Simons, Acting P. J., and Needham, J., concurred.A

A petition for a rehearing was denied March 24, 2011, and the petition of appellant Ramzy Ayyad and appellant Sprint Spectrum, L.P., for review by the Supreme Court was denied June 15, 2011, S192165. Baxter, J., and Werdegar, J., did not participate therein.