Opinion

I. Introduction

In this case we modify a $13,971,830 judgment in a class action which challenged the assessment of fees by First Interstate Bank (FICAL) against breaching credit card customers who failed to make minimum monthly payments when due or exceeded their credit limits. We reduce the judgment by $9,076,304, which represents the amount the trial court found to be net revenue or benefit from finance charges imposed by FICAL on delinquent and overlimit balances. We hold FICAL was entitled to retain such revenue because the interest provided for in the credit card agreement remained chargeable after the breach thereof. Further, such revenue cannot be treated as gain made possible only by the breaches, and thus is not an offset against FICAL’s recoverable actual damages, because FICAL could have made similar loans to other customers had the breaches never occurred. We also hold that the underlying credit card agreements invoke the provisions of Civil Code section 1671, subdivision (d), prescribing the standard of invalidity of liquidated damages in a consumer contract for services, and that substantial evidence supports the remaining $4,895,526 portion of the judgment.

II. Background

This litigation is similar to Beasley v. Wells Fargo Bank (1991) 235 Cal.App.3d 1383 [ 1 Cal.Rptr.2d 446 ], in which Division Four of this appellate district affirmed a $5,227,617 judgment in a class action challenging two types of fees imposed on credit card customers: a “late” fee assessed against customers who did not make monthly payments on time, and an “overlimit” fee charged to customers whose account balances in a given month exceeded their credit limits.

Beasley upheld a determination that the fees were invalid as liquidated damages under subdivision (d) of Civil Code section 1671, which applies to certain consumer contracts and provides that a provision for liquidated damages in such contracts “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.” Beasley held, based on Garrett v. Coast & Southern Fed. Sav. & Loan Assn. (1973) 9 Cal.3d 731 [ 108 Cal.Rptr. 845 , 511 P.2d 1197 , 63 A.L.R.3d 39 ], that “plaintiffs still remained liable for ‘the actual damages resulting from’ their late and overlimit activity, which would include ‘administrative costs reasonably related to collecting and accounting for’ late and overlimit balances.” (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at p. 1390 , quoting Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 741.) In Garrett , which concerned the invalidity of fees charged for failure to make timely installment payments on promissory notes secured by deeds of trust, the court held that the borrower “remains liable for the actual damages resulting from his default. The lender’s charges could be fairly measured by the period of time the money was wrongfully withheld plus the administrative costs reasonably related to collecting and accounting for a late payment.” (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 741.)

FICAL’s form credit card agreements authorize the late and overlimit fees. In 1982, the late fee was set at 5 percent of the amount past due, with a minimum charge of $3 and a maximum charge of $5 (before then, the minimum charge was $1), and the overlimit fee was set at $10 for outstanding balances exceeding credit limits (in practice, by more than 10 percent).

The class was certified to consist of all FICAL credit cardholders who were assessed late or overlimit fees after February 10, 1983. After a nonjury trial, the court issued a lengthy and thorough statement of decision, concluding that the challenged fees were void under Civil Code section 1671, subdivision (d), and their imposition was an unlawful business practice under Business and Professions Code section 17200 et seq. The court rendered judgment for plaintiffs in the total sum of $13,971,830 for fees paid through October 21, 1991, consisting of “$9,569,560 for late fees collected in excess of defendant’s actual costs resulting directly from late payments, and $4,402,260 for overlimit fees collected in excess of defendant’s actual costs resulting directly from overlimit charges.”

III. Discussion

A. Finance Charges as a Basis for Reducing FICAL’s Damages

In calculating FICAL’s actual damages resulting from the late payment and overlimit breaches, which damages are to be deducted from plaintiffs’ recovery, the trial court reduced damages by what the statement of decision called “the monetary benefit earned by the bank resulting from the very same breaches, i.e., interest earned by the bank on late and overlimit balances.” In other words, the court determined that FICAL mitigated its damages by imposing finance charges, at contract rates of up to 21 percent, on the portions of account balances that were delinquent or overlimit.

The court determined that during the class period, FICAL collected finance charges of $15,134,864 on delinquent balances and $1,145,544 on overlimit balances, for a total of $16,280,408. According to the court, the bank’s cost of borrowing these sums (for lending to plaintiffs), based on the average federal funds interest rate during the class period, was $6,745,554 for delinquent balances and $458,550 for overlimit balances, for a total of $7,204,104. By this reckoning, the net “benefit” to FICAL from these finance charges (the charges collected minus the bank’s cost of borrowing) was $9,076,304. This sum is, in effect, the major component of plaintiffs’ $13,971,830 judgment because as an offset against FICAL’s actual damages that reduce plaintiffs’ recovery, it becomes a part of plaintiffs’ recovery.

FICAL contends this was error because the result is to deny the bank its full contract rate of interest on delinquent and overlimit balances, to which it is entitled by statute and case authority, improperly replacing the contract rate with the federal funds interest rate.

We agree with FICAL’s contention. It is wrong to compel FICAL, in effect, to finance delinquent and overlimit balances at rates that are lower than are charged on nondelinquent and within-limit portions of accounts and to nondefaulting cardholders. This would be inconsistent with FICAL’s entitlement to actual damages “fairly measured by the period of time the money was wrongfully withheld . . . .” (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 741.)

In its statement of decision the trial court acknowledged FICAL’s entitlement to the full contract rate of interest on late and overlimit balances, but treated what it considered to be net revenue or benefit as an offset against FICAL’s damages: “The court holds that a determination of FICAL’s damages resulting from ‘money wrongfully withheld’ in the form of late payments and overlimit transactions must take into account the revenue earned by FICAL on the amounts in breach, as well as the costs incurred in borrowing those sums, i.e., the bank’s cost of funds.” Thus the $9,076,304 figure at issue in this case was arrived at by the trial court by taking the difference between its determination of the amount of interest to which the bank is entitled at the contract interest rate ($16,280,408) and the bank’s cost of federal funds ($7,204,104).

The court’s holding is in error for several reasons. First, the parties and the trial court all agree that the contract rate of interest agreed upon continues to be chargeable on the delinquent and overlimit balances after the breach. Civil Code section 3289, subdivision (a), expressly provides that “[a]ny legal rate of interest stipulated by a contract remains chargeable after a breach thereof, as before, until the contract is superseded by a verdict or other new obligation.” This proposition is so well established that both parties cite authorities in support thereof which were decided over a century ago. (Casey v. Gibbons (1902) 136 Cal. 368, 371 [ 68 P. 1032 ] [“The rate of interest was stipulated in the contract and ‘remains chargeable after a breach thereof, as before, until the contract is superseded by a verdict or other new obligation’ ”]; Kohler v. Smith (1852) 2 Cal. 597, 597-598 [“This language [in the act] is very explicit, and shows that the intention of the act was two-fold; first, that money demands after maturity should draw interest; and second, that they should draw interest at whatever rate was expressed in the written contract”].)

In the case before us, however, the trial court, in effect, reduced the contract rate of interest recoverable by FICAL. Plaintiffs say this is untrue and they point to the decision of the trial court. The court ruled as follows: “In determining the amount of excessive late fees that must be returned to class members ... the Court has used the same methodology it used with respect to the overlimit fees; i.e., the Court has determined the amount of revenue FICAL collected in late fees and finance charges on delinquent balances, and subtracted FICAL’s actual damages that may be attributed to late payments, consisting of the bank’s cost of funds [federal funds] for extending credit on delinquent balances and administrative expenses reasonably related to collecting and accounting for delinquent balances.”

Plaintiffs maintain that this shows that the court permitted FICAL to recover all interest on the delinquent and overlimit balances at the full contract rate and that the court was correct in finding that FICAL’s actual damages included its cost of federal funds for extending credit on these balances rather than interest at the contract rate, because the difference between the two rates was a benefit FICAL received as a result of the breach. The fallacy of this argument is that under the trial court ruling FICAL is given the finance charges in one breath and more than half of what it received is taken away in the next breath. FICAL is in effect being denied recovery of any contract interest on delinquent and overlimit balances over and above the federal funds interest rate.

If the rule advanced by plaintiffs and adopted by the trial court was permitted to stand it would result in an anomalous situation in commercial loan transactions. As examples: if a credit card did not have a provision for a late payment fee and a cardholder breached the agreement by failing to make a timely minimum monthly payment, the bank would be entitled to interest at the full contract rate on the delinquent balance, but under the trial court’s ruling the bank’s recovery would be reduced by the benefit received, i.e., the difference between the contract interest rate and the cost of acquiring federal funds. This would mean that delinquent cardholders would pay less on their delinquent balances than nondelinquent cardholders would pay on their nondelinquent balances. Likewise, if a cardholder in this situation was paying interest at the contract rate but failed to make a minimum monthly payment, his or her rate on the delinquent balance would immediately drop whereas his or her rate on the nondelinquent balance would remain at the contract interest rate. Additionally, if the rule adopted by the trial court applies to credit card loans there would seem to be no reason for it not applying to commercial loans generally. Thus, in this situation if a borrower was in default in making payments it would be necessary to determine the source and cost of funds available to the lender, and the lender’s actual recovery of interest on delinquent balances would be limited to the cost of funds. This would cause great confusion and uncertainty for borrowers and lenders alike, and result in time and expense in ascertaining the source and cost of funds available to the lender. This is certainly contrary to the intent of Civil Code section 3289, subdivision (a), which specifically provides for interest to continue at the rate stipulated by the contract. It is also contrary to the holding in Garrett where our Supreme Court emphasized the fixed nature of damages resulting because of the wrongful withholding of money. (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 741, fn. 11 .)

Second, plaintiffs’ theory is that breach actually conferred a benefit on FICAL by enabling the bank to extend additional credit to them, to the extent of the delinquent and overlimit portions of outstanding balances, at the contract rate of interest. By accepting this theory, the trial court treated the additional finance charges as gains that could not have been made had there been no breach, and thus an offset against FICAL’s actual damages.

Such treatment was improper. A converse rule for postbreach gains applies here: “Gains made by the injured party on other transactions after the breach are never to be deducted from the damages that are otherwise recoverable, unless such gains could not have been made, had there been no breach.” (5 Corbin on Contracts (1964) § 1041, p. 256, italics added.) That is, only those gains that could not have been made had there been no breach are treated as an offset against actual damages.

FICAL’s collection of finance charges on delinquent and overlimit balances was not such gain. The delinquent and overlimit cardholders forced FICAL to extend additional credit to them, resulting in more finance charges, but these additional loans could have been made to other customers as well, in many instances with substantially less risk of default. The crucial point is that FICAL’s credit card business is expandable: FICAL could have made similar loans to other customers had breach never occurred, and thus the gain from finance charges on delinquent and overlimit balances was not made possible only by breach.

Two cases from New Jersey and California, involving breaches of contracts to lease jukeboxes, illustrate this concept of expandability.

In the New Jersey case, Locks v. Wade (1955) 36 N.J.Super. 128 [ 114 A.2d 875 ], defendant lessee repudiated a contract to lease a jukebox for two years, and the trial court awarded plaintiff lessor the amount that would have been paid under the lease. On appeal, defendant asked for credit in an amount plaintiff had received by leasing the component parts of the jukebox to other persons. The appellate court rejected this request because the jukebox was “an article of which the supply in the market is for practical purposes not limited . . . .” (Id. at p. 876.) Because the jukebox, unlike real property, was not unique but duplicable, “. . . 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.” (Ibid.) Thus, the court concluded, “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.” (Id. at p. 877.) Because gain from postbreach leases could have been made even without the breach, such gain did not reduce the lessor’s damages.

The New Hampshire Supreme Court later explained the rule of Locks v. Wade as applying where plaintiff sells a product “that can be produced according to demand. The theory is that the second sale would have occurred even if the defendant did not breach his contract." (Kearsarge Computer, Inc. v. Acme Staple Co. (1976) 116 N.H. 705 [ 366 A.2d 467, 471 , 86 A.L.R.3d 1081 ].) Certain businesses “are deemed to be expandable. The law presumes that they can accept a virtually unlimited amount of business so that income generated from accounts acquired after the breach does not mitigate the plaintiff’s damages.” (Ibid.)

In the California case, Seaboard Music Co. v. Germano (1972) 24 Cal.App.3d 618 [ 101 Cal.Rptr. 255 ], upon the sale of a tavern the seller repudiated his contract to lease a jukebox and a pool table, doing so on the advice of business opportunity salespersons who arranged the sale. The trial court ordered the seller and salespersons to pay the amount remaining due under the lease. On appeal, they claimed the lessor should have been required to mitigate damages by re-leasing the equipment elsewhere. (Id. at pp. 621-622.) As in Locks v. Wade , the Court of Appeal rejected this argument because, due to the expandable nature of the lessor’s business, the equipment could have been leased elsewhere even without the breach: “The evidence here shows plaintiff was engaged in the business of leasing coin-operated equipment; it had a warehouse full of equipment similar to that leased to [the lessee] from which it serviced its customers and from which it could service any additional leases negotiated, irrespective of whether [the lessee] fulfilled or breached his obligation.” (Id. at p. 623.) Again, because gain from postbreach leases could have been made even without the breach, the possibility of such gain did not reduce the lessor’s damages.

In the present case, plaintiffs treat credit card financing as something that can be produced according to demand: plaintiffs’ position is that the bank meets any demand for additional credit simply by borrowing federal funds. But on this theory, additional credit card financing using federal funds is like the warehouse full of jukeboxes in Seaboard: there are enough jukeboxes to supply any additional leasing demands, and there are enough federal funds to supply any additional financing demands, irrespective of any breach of an existing contract. Because the credit card business, like the jukebox leasing business, is expandable, FICAL could have loaned additional money elsewhere to nondefaulting customers in amounts equivalent to the delinquent and overlimit sums, even if the delinquency and overlimit breaches had never occurred. Thus, in both situations, gain from the postbreach transactions is not gain that could not have been made had there been no breach, and is not an offset against the aggrieved party’s actual damages.

In short, plaintiffs’ delinquency and overlimit breaches did not confer a benefit made possible only by the breaches. FICAL collected finance charges on the delinquent and overlimit balances, but could have made similar loans to other customers at contract rates up to 21 percent even if the breaches had never occurred. Because there is no gain that could not have been made had there been no breach, the finance charges are not an offset against FICAL’s actual damages, and thus are not a proper component of plaintiffs’ recovery. The judgment must be reduced by $9,076,304.

Third, the trial court held that the plaintiffs’ breach of the agreement conferred a benefit on FICAL because FICAL earned interest on the late and overlimit balances. The court ruled that FICAL was entitled to continue to impose the contract rate of interest on late and overlimit balances during the period of breach as compensation for amounts wrongfully withheld but the benefit received thereby had to be returned to class members. The court found that the benefit conferred on FICAL was the difference between the interest earned at the contract interest rate and the cost of funds acquired at the federal funds rate.

As we have already set forth above, the methodology used by the trial court is in error, but we note in passing that even under the trial court’s own approach, its determination of benefit is flawed. When it is said that the bank is entitled to continue to earn interest at the contract rate on delinquent and overlimit balances, it cannot be said that the cost of funds is the only direct cost incurred in earning this interest. It is important to note that we are not talking here about administrative costs related to collecting and accounting for delinquent or overlimit balances. The trial court in its decision uses two separate sets of costs. One is a cost of extending credit and the other is the cost of administrative expenses reasonably related to collecting and accounting for delinquent and overlimit balances. On the point we are discussing here, we are concerned with the former. In its statement of decision the court said: “The amount of [the] benefit to the bank was, of course, dependent upon the bank’s cost of extending credit for ‘the period of time the money was wrongfully withheld.’” In making an allocation of costs the court describes in detail the cost of extending credit and the cost of administrative expenses reasonably related to collecting and accounting for delinquent and overlimit balances. In determining the cost of extending credit on delinquent and overlimit balances, however, the trial court used only the cost of funds at the federal funds interest rate. Thus, even under the trial court’s own methodology, this determination is in error. If there is to be a finding of a benefit conferred upon FICAL, then at least a portion of the costs directly related to extending credit, in addition to the cost of federal funds, must be taken into account in determining the amount of the benefit.

B. Application of Civil Code Section 1671, Subdivision (d)

Civil Code section 1671 prescribes two alternative standards for determining the validity of a liquidated damages provision. Under subdivision (b), the provision “is valid unless the party seeking to invalidate the provision establishes that the provision was unreasonable under the circumstances existing at the time the contract was made.” Under subdivision (d), the provision “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.”

Subdivision (c) of Civil Code section 1671 prescribes when subdivision (d), rather than subdivision (b), is applicable:

Thus, under subdivision (c)(1) of Civil Code section 1671, the standard prescribed by subdivision (d) applies to certain consumer contracts. The trial court ruled that subdivision (d) applies here because a credit card agreement is a consumer contract for the provision of “services.” FICAL challenges this ruling, arguing that the credit card agreement is a contract for the extension of credit rather than services, and legislative history indicates an extension of credit is not a consumer contract within the meaning of subdivision (c)(1).

This argument overlooks the dual nature of a credit card agreement. We need not decide whether an extension of credit is a consumer contract within the meaning of Civil Code section 1671, subdivision (c)(1), because a credit card agreement is much more than that, encompassing convenience services in addition to extension of credit.

A textbook on commercial banking explains these two discrete functions: “The popularity of credit cards is due to the many advantages they offer as a means of payment. These advantages have created two general distinct patterns of credit card use among cardholders—convenience and revolving credit. Many cardholders pay their outstanding balances in full each month; consequently, they incur no monthly finance charge. In fact, nearly half of the cardholders can be classified as convenience users. The remaining cardholders use credit cards as a source of credit and infrequently pay their entire outstanding monthly balance. Both of these uses have distinct advantages over cash, checks, and other means of payment. Convenience use minimizes the need to carry cash, allows the user to defer payment for goods and services for a short time, and establishes a favorable payment record that is important in credit evaluations. Revolving credit users realize the same advantages plus one other, namely, they increase their ability to purchase goods and services and in so doing avoid the red tape involved in obtaining a personal loan. Moreover, the credit card holder has considerable flexibility in the timing and amount of debt repayment.” (Reed & Gill, Commercial Banking, supra, p. 337, italics added.)

An economist whose work is cited by amicus curiae California Bankers Association similarly describes credit cards as encompassing two features: “payments services” for “convenience users” who wish to make purchases “without paying cash or writing a check,” and “credit features” for those who wish to borrow. (Litan, The Economics of Credit Cards, supra, pp. 2, 4.)

The convenience feature of credit cards is surely a “service” within the meaning of Civil Code section 1671, subdivision (c), wholly apart from the credit feature. Observers of the banking industry view the convenience feature as such; the publications quoted above both include references to “credit card services.” (Reed & Gill, Commercial Banking, supra, at pp. 339-340; Litan, The Economics of Credit Cards, supra, at p. 2.) A credit card user enjoys various benefits other than borrowing—primarily cashless and checkless purchasing—regardless of whether the credit feature is used. Indeed, convenience use without borrowing is the “reason that some banks levy a flat charge on the use of the card." (Reed & Gill, supra, at p. 339.) Thus, some users even pay for these two features separately: their annual charge for the card is attributable to the convenience feature, while they pay for use of the credit feature through finance charges.

FICAL argues that late and overlimit fees concern only the extension of credit and have nothing to do with other services. But that is untrue: a cardholder can accidentally exceed a credit limit or make a full payment a few days late and thus incur a fee without ever having intended to borrow. Going overlimit and paying late are as much a breach of the agreement for convenience services as a breach of the agreement for credit—the bank provides both features in exchange for timely payment and use of the card within the credit limit.

In any event, all cardholders receive the convenience service, even where there is also use of credit. Thus, in all instances, the credit card agreement is in part a “contract for the retail purchase” of “services.” (Civ. Code, § 1671, subd. (c)(1).) As such, it invokes subdivision (d) of Civil Code section 1671, because “the liquidated damages are sought to be recovered from” a party to such a contract. (Civ. Code, § 1671, subd. (c).)

This was in fact the primary basis for the court’s determination that Civil Code section 1671, subdivision (d) applies. The court concluded in its statement of decision that the credit card agreement “is a contract between a consumer and the bank for the provision of a service that allows the consumer to use the credit card as a common convenience, a form of payment that can be used in any number of transactions as the medium of exchange without the requirement of cash currency.” The court ruled properly on this point.

C. Application of the Reasonable Endeavor Test

For liquidated damages to be valid under subdivision (d) of Civil Code section 1671, it must have been “impracticable or extremely difficult to fix the actual damage.” (Civ. Code, § 1671, subd. (d).) Further, 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.” (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 739 .) Absent either of these elements, a liquidated damages provision is void, although breaching parties remain liable for the actual damages resulting from the breach. (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at p. 1390 .) FICAL contends the evidence demonstrated a “reasonable endeavor” as a matter of law and the court misapplied the “impracticability” test.

We first address the reasonable endeavor issue. According to Garrett , whether liquidated damages are the result of a reasonable endeavor to estimate fair compensation for loss “depends upon the motivation and purpose in imposing such charges and their effect.” (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 740.) If a charge “is designed to exceed substantially the damages suffered . . . , the provision for the additional sum, whatever its label, is an invalid attempt to impose a penalty . . . .” (Ibid.) Thus, Garrett instructs us that the focus is not, as FICAL claims, 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.

The evidence was in conflict as to FICAL’s motivation and purpose. On the one hand, David Garrett, FICAL’s former vice-president who made the recommendation in 1982 to impose the overlimit fee and change the late fee, testified at trial that the purpose of the fees was “to recoup our costs” and not to penalize breaching cardholders.

On the other hand, there was documentary and testimonial evidence indicating that the purpose of the fees was to generate profit, and thus to collect charges “designed to exceed substantially the damages suffered.” ( 9 Cal.3d at p. 740 .) An internal bank memorandum by Garrett dated July 8, 1982, projected “new revenue” of $250,000 from overlimit fees in 1982, without any mention of costs. Garrett testified on cross-examination that at the time he wrote this memorandum, “I expected the imposition of those fees to generate new revenue that overall would increase the profitability of the credit card business.” In deposition testimony read into the record at trial, Garrett testified that he had considered late fees to be “a good source of revenue,” and he had expected the bank to “make a profit” from overlimit fees. He explained at trial that the bank was looking for new sources of revenue in 1982 because, among other things, interest rates nationwide were high, approaching the rate charged to FICAL’s credit card users. Garrett admitted (and it is undisputed) that when FICAL made the decision to impose the overlimit fee and change the late fee, the bank had conducted no study or analysis of the costs resulting from late and overlimit activity.

This evidence was the basis for the trial court’s determination that there was no reasonable endeavor to estimate fair compensation for loss. The court explained in the statement of decision, “This evidence leads the Court to the inescapable conclusion that FICAL imposed the late and overlimit fees in 1982 considering the additional revenue they would generate, rather than the actual losses caused by cardholder breach. FICAL never intended, in 1982 or any subsequent time, for the late and overlimit fees to approximate actual loss or damage; the fees were intended to generate revenue.”

From an appellate perspective, whether the bank’s motivation and purpose was to generate profit or to recoup losses is purely a substantial evidence question turning on Garrett’s credibility. Was Garrett to be believed when he claimed on direct examination that FICAL’s purpose was to recoup losses? Or should greater credence have been given to his concessions on cross-examination and in deposition testimony regarding his anticipation of new revenue and increased profit at a time of high interest rates, as well as to the inference to be drawn from the lack of any cost study or analysis that the bank’s focus was not on recouping losses? This credibility issue lies within the exclusive province of the trial court, and is not subject to independent determination on appeal. (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at p. 1395 .) Under the most basic rule of appellate review, we are not permitted to second-guess the trial court’s factual determination as to the bank’s motivation and purpose, but are bound by that determination. (See, e.g., Bowers v. Bernards (1984) 150 Cal.App.3d 870, 873-874 [ 197 Cal.Rptr. 925 ].)

FICAL contends, however, that the evidence demonstrated a reasonable endeavor as a matter of law. The bank relies on Garrett’s testimony, “I felt that I 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 FICAL made a profit from late and overlimit fees. The trial court said in the statement of decision that Garrett’s “general knowledge” and “opinion” on costs “simply do not constitute the type of reasonable endeavor to estimate actual loss or damages that is required . . . .” According to FICAL, this was a legal conclusion, not a credibility determination, and thus is subject to de novo appellate review.

If this was a legal conclusion, it was correct. The focus is on FICAL’s motivation and purpose, not whether Garrett personally had a “good understanding” of costs. If, as the trial court found in evaluating conflicting evidence, FICAL’s purpose was to generate revenue (i.e., profit) rather than provide compensation for loss, then Garrett’s general knowledge of costs cannot transform the fees into the result of a reasonable endeavor to provide compensation. The pivotal factor is the bank’s purpose in imposing the fees, not Garrett’s personal knowledge of costs.

The court properly evaluated Garrett’s personal knowledge of costs within the context of the factual issue presented as to FICAL’s motivation and purpose. The court observed in the statement of decision, “Given the accounting precision and care usually expected from those in the banking profession, the Court finds it incredible that, had the bank in fact intended to recover its actual anticipated damages by imposing the fees, it would have adopted the fee provision without any concrete basis for Mr. Garrett’s opinion and without the benefit of a cost study or other appropriate analysis to determine its damages resulting from cardholders’ breach of contract.” Assuming Garrett had the “good understanding of our costs” that he claimed, the court nevertheless drew an inference, from FICAL’s failure to verify Garrett’s “good understanding” with a cost study or some other analysis, that the bank was looking for revenue enhancement, not compensation for loss. Again, on appeal we are not permitted to draw a contrary inference. (Bowers v. Bernards, supra, 150 Cal.App.3d at p. 874 .)

FICAL claims the trial court erred by requiring a cost study, because no published opinion makes such a study essential to a valid liquidated damages provision. But the court did not require a cost study per se; the statement of decision merely said there had to be some form of analysis to determine loss: “While the Court declines to specify, as a matter of law, what particular type of analysis is required, in order to establish the reasonable endeavor required by Garrett and Beasley , a bank is required to show that it actually engaged in some form of analysis to determine what losses it would sustain from breach, and that it made a genuine and non-pretextual effort to estimate a fair average compensation for the losses to be sustained.” (Original italics.) In the present context, the court stated the law correctly. There must be a reasonable endeavor “ ‘to estimate a fair average compensation for any loss that may be sustained’ ” (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 739), and such an estimate cannot occur without some sort of analysis of the loss that is to be compensated.

FICAL also contends the court erred in ruling that certain cost analyses by the bank in 1988 (so-called “unit cost grid” studies commenced in 1986) and in 1989 (a Price Waterhouse cost accounting study commissioned in 1987 in response to this litigation) were irrelevant to the reasonable endeavor issue. According to the statement of decision, 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, FICAL’s knowledge of costs in 1988 and 1989 is simply irrelevant to any determination of validity.” Again, this was a correct statement of the law. The “amount” of liquidated damages “must represent the result of a reasonable endeavor” to estimate fair compensation. ( 9 Cal.3d at p. 739 , italics added.) 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. Analysis of costs in 1988 and 1989 is not pertinent to FICAL’s motivation and purpose when it decided on the amounts of the late and overlimit fees in 1982.

We find no error in the court’s ruling on the reasonable endeavor issue. Consequently, we need not decide FICAL’s claim of error on the impracticability issue, since liquidated damages are void if either of those elements is absent. (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at p. 1390 .)

The rest of this opinion addresses various other substantial evidence issues.

D. Sufficiency of Testimony on Actual Damages

IV. Conclusion

We find one error of considerable dimension, the court’s subtraction of net additional finance charges from FICAL’s actual damages from breach, which requires us to reduce the amount of the judgment by $9,076,304. None of FICAL’s other claims of error undermine the remaining $4,895,526 of the $13,971,830 judgment.

This case would appear to close a chapter in credit card fee litigation. Effective January 1, 1995, legislation expressly permits credit card issuers to impose late fees of $7 to $15 and overlimit fees of $10. (Fin. Code, § 4001, subd. (a); see Civ. Code, § 1671, subd. (a) [this statute inapplicable where other statute prescribes rule for validity of liquidated damages].)

V. Disposition

The judgment is modified to award plaintiffs a total sum of $4,895,526, and is affirmed as modified. The parties shall bear their own costs on appeal.

Chin, P. J., and Corrigan, J., concurred.

A petition for a rehearing was denied October 10, 1995, and appellant’s petition for review by the Supreme Court was denied December 14, 1995.

The California Supreme Court recently held that the term “interest” in a provision of the National Bank Act should be construed to cover late payments, if such fees are allowed by a national bank’s home state. (Smiley v. Citibank (1995) 11 Cal.4th 138 [ 44 Cal.Rptr.2d 441 , 900 P.2d 690 ].) In that case the defendant bank’s home state was South Dakota, and thus, federal law preempted California from applying its own state law. That decision is inapposite to the case before us. FICAL’s home state is California and California law applies. “Federal funds are deposit balances held with Federal Reserve banks. A commercial bank may have excess balances because of an unexpected inflow of deposits or a decline in loans. Since these funds are noneaming assets, banks are willing to make them available to other banks for a short period of time, and those that need funds to comply with reserve requirements or to purchase assets are willing to purchase such balances. . . . Today the large commercial banks view these funds as simply an alternative source that may be obtained at a price to meet liquidity and loan needs.” (Reed & Gill, Commercial Banking (4th ed. 1989) p. 150.) This was not a component of the judgment in Beasley . California Bankers Association’s brief as amicus curiae includes an economist’s assertion that cardholders who have been assessed a late fee are responsible for credit losses at a rate four to six times higher than for all accounts generally. (Litan, The Economics of Credit Cards (1993) pp. 7-8.) Because we reduce the judgment by the entire amount of the finance charge component, we need not address FICAL’s arguments that the trial court erred in calculating that amount. This issue was not presented in Beasley , where the bank “expressly declined to challenge the application of section 1671, subdivision (d) to this case, effectively conceding that this is a ‘consumer’ action.” (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at p. 1399, fn. 7 .) In asserting that section 1671, subdivision (c) does not encompass the extension of credit as a consumer contract, FICAL relies on a comparison of similar language in the definition of “consumer” in the Consumer Legal Remedies Act (Civ. Code, § 1761, subd. (d)), where the Legislature deleted purchasers of “credit” from the initial proposed definition (Assem. Bill No. 292 (1970 Reg. Sess. (Jan. 21, 1970, & as amended Aug. 7, 1970)), with the definition of “consumer” in the Consumer Affairs Act (Bus. & Prof. Code, § 302, subd. (c)), which includes purchasers of “credit.” According to FICAL, this indicates the Legislature intended to exclude “credit” contracts from the scope of the Consumer Legal Remedies Act (see, e.g., Wilson v. City of Laguna Beach (1992) 6 Cal.App.4th 543, 555 [ 7 Cal.Rptr.2d 848 ] [rejection of provision from act as originally introduced is “most persuasive” that act should not be interpreted to include what was left out]), and thus the similar language in subdivision (c) of section 1671 should likewise be construed to exclude credit agreements from the scope of its definition of consumer contracts. FICAL began charging a $15 annual membership fee in 1982. For example, a witness for plaintiffs testified that she inadvertently forgot to pay a $6.50 balance due on June 22, 1987, made the payment by July 8, 1987, but was nevertheless charged a $3 late fee. FICAL also cites Peterson v. Wells Fargo Bank (N.D.Cal. 1981) 556 F.Supp. 1100, 1111 , which held, in a different context, that a credit card issuer is not a “retail seller” of goods or services within the meaning of California law regulating finance charges on retail installment sales (Civ. Code, § 1801 et seq.) because the issuing bank “has no interest in any particular product." ( 556 F.Supp. at p. 1111 ) In the present context, however, as between the bank and cardholder there is a particular service in which the bank has an interest, the convenience service. As for Garrett’s claim of unprofitability, the court found otherwise. The court also ruled in the statement of decision that FICAL did not use the unit cost grids or Price Waterhouse study “to perform the kind of analysis required by Garrett and Beasley”—i.e., there was no motivation and purpose of estimating fair compensation at those times with the fee amounts previously set in 1982. Again, this is a factual determination, which FICAL has shown no basis for disturbing. Indeed, a former FICAL vice-president who became responsible for preparing unit cost grids in 1988 testified that their focus was not on estimating damages from late and overlimit activity. There was similar testimony that the Price Waterhouse study was not performed to determine actual damages from breach. Thus, the post-1988 analyses cannot, as FICAL claims, validate fees charged after 1988. There is merit, however, in FICAL’s contention that the trial court misapplied the “impracticability” element by finding in the statement of decision that “it would have been neither difficult nor impracticable for FICAL to have estimated its average damage resulting from cardholder breach.” (Italics added.) The test prescribed by subdivision (d) of Civil Code section 1671 is whether “it would be impracticable or extremely difficult to fix the actual damage.” (Italics added.) The proper focus is on actual damage, not average damage. Use of average damage in the test of impracticability seems inconsistent with the reasonable endeavor test, which requires an attempt to estimate a fair average compensation for loss. (Garrett v. Coast & Southern Fed. Sav. & Loan Assn., supra, 9 Cal.3d at p. 739 .) How can we logically require, at the same time, impracticability of fixing average damage yet a reasonable endeavor to estimate average compensation? The point is inconsequential, however, since the element of reasonable endeavor was not established in this case. See footnote, ante, page 274. Our judgment leaves intact the trial court’s determination that the fees were unlawful under Civil Code section 1671. Thus, we necessarily reject FICAL’s argument that, for want of violation of section 1671, imposition of the fees was not an unlawful business practice under Business and Professions Code section 17200 et seq. (See Bondanza v. Peninsula Hospital & Medical Center (1979) 23 Cal.3d 260, 265-267 [ 152 Cal.Rptr. 446 , 590 P.2d 22 ].)