The opinion
WILKEY, Circuit Judge,
dissenting in part:
While I concur in the bulk of the majority’s opinion, I believe my colleagues too hastily accept one part of the Commission’s ratemaking prescriptions in Docket No. 18,-128, in re American Telephone & Telegraph Co., 1 which I find arbitrary and capricious. The FCC’s order in Docket No. 18,128 prescribes the use of Fully Distributed Costs (FDC) as the required standard for calculating rates both in monopoly markets and in competitive markets. As a ceiling for rates in monopoly markets, it is reasonable to use a standard based on FDC as a means to prevent monopoly pricing. But for the entirely different task of setting a rate floor in competitive markets, FDC pricing should not be hastily approved without scrutiny of its economic rationality in this context. In its choice of FDC pricing in this case, the FCC has shown a serious misunderstanding of the nature of pricing decisions in competitive markets. This misunderstanding has led the Commission to adopt FDC pricing without rational consideration of its anti-competitive effects in a competitive market, and without showing that this standard is rationally related to the Commission’s legitimate regulatory purposes.
I must therefore dissent from Part II A of the majority opinion insofar as it upholds FDC as a standard for establishing rates in competitive markets. For the same reason, I must also concur separately in the majority’s Part II C; while I agree in vacating and remanding the Commission’s findings as to specific rates, I would not permit the Commission on remand to evaluate specific rat§s by a standard that embraces FDC for minimum pricing in competitive markets.
As the majority opinion states, our task in reviewing challenged Commission actions of this sort is limited to keeping the Commission within proper bounds for the reasonable exercise of its discretion. While agencies do have broad discretion in rate-making proceedings, 2 we must also keep in mind that “the width of administrative authority must be measured in part by the purposes for which it was conferred . .” 3 The rationality of the Commission’s prescribed FDC costing methodology is challenged both by petitioner American Telephone and Telegraph Co. (AT&T), 4 and by the United States Department of Justice in its role as a statutory respondent. 5 The significance of the fact that the Antitrust Division and AT&T — not usually found as allies in legal contests — are in thorough agreement that the Commission has embraced some fundamentally unsound economics, as the underpinning of its whole decision on this point, should not be minimized.
To evaluate the rationality of the Commission’s prescribed FDC pricing for competitive markets, we must therefore first define the legitimate purposes of FCC rate regulation. The Commission derives its authority to regulate rates for communication services from Title II of the Communications Act of 1934. The Act sets broad standards requiring that rates be “just and rea *269 sonable” 6 and prohibiting “unjust or unreasonable discrimination” and “undue or unreasonable preference or advantage.” 7 In monopolized markets FCC rate regulation serves the purpose of limiting carriers to just and reasonable rates in order to prevent monopoly pricing. In competitive markets, on the other hand, just and reasonable prices are normally maintained through the effect of free competition — indeed, regulation of “just and reasonable” rates is essentially a replacement or surrogate for the effects of free competition. 8
As a general matter, then, competitive markets do not need the constant rate regulation that is necessary to ensure just and reasonable prices in monopoly markets. But a potential does exist for unjust or discriminatory prices to arise in non-monopoly markets, in a manner that can warrant rate regulation. When a firm operates in a regulated market where it enjoys a monopoly, and also operates in a separate market where it faces competitors, there is a danger that the firm may reduce its rates in the competitive market below the level that would fully compensate for its costs of operation in that market. The use of this technique to drive competitors out of the market or to keep potential competitors out is often called “predatory pricing.” 9 If a regulated firm such as AT&T engaged in predatory pricing, its losses or below-normal profits in the low-price competitive market could be offset by increasing its profits in the firm’s monopoly markets. Even if the firm’s overall rate of return were within just and reasonable limits set by the regulatory agency, the customers in the monopolized market could find themselves paying higher than reasonable rates in order to finance the firm’s predatory pricing activities in the competitive market. An unjust and discriminatory system of “cross-subsidies” could thus arise, with the monopoly market customers paying more than their fair share in order to subsidize the firm’s activities in competitive markets.
The FCC focuses on this danger to justify its regulation of minimum rates for AT&T in the competitive market of private line services. As its basic regulatory objective here, the FCC seeks costing and pricing standards that will allow AT&T to compete in a new and competitive market “without allowing it to use revenues from captive monopoly markets as a source for cross-subsidies.” 10 The existence of cross-subsidies would violate the Commission’s goal of equitable and nondiscriminatory treatment of all service users, 11 and would permit AT&T to exploit resources acquired through its privileged monopoly status in other markets to gain an unfair advantage against other firms in the competitive market. 12
There is no question that preventing cross-subsidies is a legitimate regulatory purpose of the FCC; the issue is whether FDC pricing is a means rationally related to this end, or whether Long Run Incremental Cost (LRIC) pricing will more surely achieve this regulatory end. As a preliminary matter, it is important to note that prevention of cross-subsidies does not justify any and all standards for minimum rates. The establishment of exorbitantly high minimum rates would certainly prevent cross-subsidies; but it would also stifle competition and seriously harm AT&T’s consumers in the competitive market, who have a strong interest in low prices. This indicates that the goal of preventing cross-subsidization exists in tension with the goal of promoting competition. The FCC has discretion to choose within a range of minimum rate standards that are rationally re *270 lated to preventing cross-subsidization. But if the Commission sets minimum rates at a level higher than is rationally related to the goal of preventing cross-subsidies, then the Commission impairs competition without rational justification, and is subject to judicial reversal for arbitrary and capricious action. This standard of review is nothing more than a specific application of the general principle quoted above, that the agency’s authority is to be measured by its purposes.
To decide what minimum level of rates is reasonable in a competitive market, the starting point is the ascertainment of the firm’s costs for providing particular services in that market. 13 Although it acknowledges that cost is not the sole criterion of reasonableness, the FCC requires as a first step that rate levels be justified in terms of the cost of providing a particular service. 14 Controversy arises from the various methods of determining costs, and from the fact that certain costs may be relevant for one purpose but not for another. The Commission maintains that in order to prevent cross-subsidization, rate levels must be calculated in relation to Fully Distributed Costs. AT&T’s proposal to calculate rates in competitive markets based on long run incremental costs, the Commission concludes, is inadequate to prevent cross-subsidies.
FDC and LRIC are alternative methodologies for calculating the costs that relate to a particular service. Under either methodology the Commission calculates the relevant costs, and then compares the cost figures with the rate level for that service. If the rates for a particular service are either higher or lower than the relevant cost figures, then the Commission can disapprove them as not cost-justified. In the present case the Commission aims to discover whether rates are sufficiently high to compensate for all relevant costs; if they are not, then the low-rate service is presumably being cross-subsidized by revenues from other services. A common aspect of both FDC and LRIC as used in this case is that a cost of capital figure is included in costs, so that when rates equal costs for a particular service, the utility is earning a rate of return equal to the cost of capital figure, in other words, a “normal” profit.
The difference between FDC and LRIC arises from the fact that some costs of any firm are inherently unattributable to any particular service category. Firms such as AT&T provide monopoly and competitive services through an integrated network with extensive use of common facilities, management, and marketing resources. 15 These common or joint costs 16 cannot be said to be related in an ascertainable way to any particular service category; rather they relate to all service categories indistinguishably. In terms of the present case, . these unattributable fixed overhead costs do not increase when AT&T decides to start offering private line services as a new category of service, and they do not decrease when ÁT&T decides to discontinue private line services.
The LRIC method assigns to each service category only those costs attributable to that category, that is, only the long-run increment to pre-existing firm costs that arises when the service category is added to existing service categories. This increment includes the variable costs incurred in providing the service plus any additional fixed overhead costs required in the long run due to the addition of the new service. But the increment does not include common or joint costs that existed before the new service was offered and would exist if the service were discontinued. Under LRIC, such unattributable costs are not assigned to any individual service category.
Under FDC pricing, on the other hand, these unattributable costs are assigned by a formula to individual service categories. *271 The formula attempts to allocate common and joint costs among the services according to the proportionate use each service makes of those facilities. Rate levels are then computed to cover the attributable costs of the service category plus its calculated share of common unattributable costs. There is of course no “right” way to assign unattributable costs — seven different formulas were considered by the Commission in this proceeding, each having various advantages and disadvantages. But FDC pricing starts from the premise that all costs must be assigned to some service category, and proceeds to assign them by a selected formula that is consistent, even if inevitably arbitrary. When calculating maximum allowable rate levels for a regulated firm, some formula must be chosen for allocating common and joint costs; otherwise such costs would be left out of the firm’s rate base, and its total revenue from all service categories would fail to cover its unattributable costs. 17
There is no serious doubt that FDC pricing is appropriate for establishing maximum allowable rate levels; for this purpose an FDC standard ensures that the firm will recover all costs while not making more than a normal allowable profit. The controversial issue is whether FDC is also reasonable for the purpose of setting minimum rates in competitive markets.
In its decision choosing FDC pricing for AT&T’s competitive markets, the FCC relies on an assumption, which runs either tacitly or explicitly throughout its opinion, that unless competitive service customers pay for the fully distributed costs of their services, monopoly service customers will be subsidizing the competitive service market. The Commission’s rejection of LRIC pricing flows from a belief that any pricing system based on marginal or incremental costs, omitting common and joint costs, will lead to cross-subsidies. This fundamental hostility toward marginal cost pricing is found most explicitly in the words of the Recommended Decision, quoted in the FCC’s opinion, that it “has not been conclusively shown that marginal cost pricing can be applied to one service without a subsidy from some other service or source.” 18
The Commission also makes this assumption explicit when, in evaluating AT&T’s rate structure, it reasons that the failure of AT&T to generate revenues for its TEL-PAK and private line services sufficient to cover fully distributed costs means that the rate structure is maintained through cross-subsidization. 19 The Commission states that “our solution in this proceeding is designed to achieve return levels for these services whereby the earnings for each competitive service can sustain that service’s viability without relying on cross-subsidization.” 20 The same assumption is stated clearly in the additional concurring views of Chairman Wiley, that under an incremental cost methodology, “AT&T would be free to subsidize private line service offerings at the expense of the telephone ratepayer.” 21
The FCC position rests on the notion that unattributable common and joint costs must be borne proportionately by monopoly and competitive service customers. An LRIC standard for minimum pricing in competitive markets, of course, does not require competitive service customers to bear any common or joint costs, but only the incremental costs the firm incurs in serving them. As stated in its appellate brief, the FCC position maintains that LRIC pricing “should not be applied selectively to AT&T’s competitive services in a manner that would burden monopoly service customers with all residual costs.” 22
The central point in this case is that, contrary to the FCC’s assumption, the long run incremental cost standard when properly applied does NOT and can NOT result in *272 a “subsidy” of competitive service customers by monopoly service customers. This point is apparent from an application of common sense to the facts of this case, and it is borne out by the analysis of economists with much greater consensus than is commonly seen in that profession. Unfortunately the majority opinion simply defers to the contrary assumption of the FCC without analyzing its rationality. 22a I believe that upon scrutiny the FCC’s assumption collapses, and leaves the Commission’s decision without rational support.
We begin from the fact that the competitive private line services market is a new market for AT&T, and that AT&T’s success in the market depends on its ability to set prices as low as its competitors. It follows from the nature of a competitive market that if AT&T raises prices it will tend to lose customers. Private line customers might either go to a competing supplier or set up their own microwave communications system. The FCC mentions that AT&T faced competition from terrestrial and satellite carriers in the intercity private line market, and from the authorization .of private microwave systems. 23 The higher AT&T is forced to maintain its prices the greater will be its loss of customers to these alternatives.
By contrast, AT&T does not face the same pressures to lower prices in its monopoly service markets. By definition, these markets do not present AT&T with competitors bidding for the patronage of various AT&T customers. Thus AT&T can keep its prices at higher levels in monopoly markets with less tendency to lose customers.
With these facts in mind, it becomes critica/ to recognize the additional fact that AT&T’s common or joint unattributable costs will exist whether or not it offers services in the competitive market. These costs existed, and were borne by AT&T’s monopoly service customers before AT&T entered the competitive market, and would again be borne fully by them if AT&T were forced out of the competitive market.
When AT&T considers whether to enter or to expand sales in a competitive market, the old monopoly service customers stand to benefit so long as the new customers bear any part of the common or joint costs. To determine whether monopoly customers will benefit from the firm’s operations in the competitive market, one need only calculate whether the revenues from the new competitive market operations pay fully for the incremental or additional costs the firm incurs for these operations. If revenues cover these costs (including cost of capital) as measured by LRIC or any similar variant of marginal cost measurement, then ANY additional revenue earned above the LRIC level is a bonus for the monopoly customers. Any revenue above the LRIC level can be used to cover part of the common or joint costs of the overall firm opérations, thus reducing the amount of common or joint costs borne by monopoly service customers and allowing a rate reduction for them. At any rate level above LRIC in the competitive market, monopoly éustomers are better off than they would be if the firm were not making such sales in the competitive market. Only if services are offered in the competitive market at less than LRIC do the monopoly service customers have to pay a higher rate than they would absent operations in the competitive market.
To say that LRIC forces all the joint or common costs onto monopoly service customers ignores the fact that they were burdened with the full measure of these costs in the first place, and the company’s expansion into the competitive market at rates above LRIC reduces the absolute amount of that burden. If monopoly service customers are not burdened more severely under *273 LRIC than they were before AT&T entered the competitive market, then there is no reasonable argument that monopoly service customers are providing a “cross-subsidy.”
Of course, the monopoly service customers would be even better off if the firm could raise rates for its competitive market services to the FDC level rather than just to the LRIC level — but only if the increase in rates does not produce a serious drop in sales. If demand is elastic, the increase in rates will actually produce a decrease in revenues. 24 The exact degree of elasticity is difficult to measure, but the general tendency is for more competitive markets to have more elastic demand. Indeed, one economist states that “the most powerful determinant of the elasticity of demand for the services of any single company is the presence or absence of competing suppliers.” 25
It is primarily a business decision, to be left to the judgment of the firm, what level of rates will produce the highest revenues. Neither an agency nor a court can predict the revenue effects of each rate increase or decrease. But so long as rates are above LRIC, agencies and courts can rest assured that monopoly service customers are NOT subsidizing the firm’s operations in the competitive market; those customers pay no more under such circumstances than they would if the firm were not in the competitive market. How much above LFIC the firm can set its prices and still remain competitive and maximize revenues, is a decision that can safely be left to private business judgment. In its pursuit of maximum allowable profit, the firm has ample incentive to set its rates as far above LRIC as the market will bear. If the firm chooses not to set its rates significantly above LRIC, the logical explanation is that competitive pressures prevent higher rates. The role of agencies and courts is only to keep prices above predatory pricing levels, to prevent cross-subsidies, and to keep rates below monopoly pricing levels. Within these bounds, pricing discretion should rest with the firm:
As long as a public utility company can take business away from its competitors at rates that cover long-run incremental costs for that business, both efficiency in the performance of the public utility function and the interest of all rate-payers recommend its being permitted to do so, all other things being equal. 26
This common sense analysis is exactly the conclusion drawn by economists. It is the nearly universal view among economists that marginal or incremental costs are more relevant to competitive market pricing than are full costs. 27 Economists generally agree that economic efficiency is served when prices approximate marginal costs. 28
For example, Aldred E. Kahn (currently head of the Council on Wage and Price Stability) states explicitly that when prices *274 for elastic demand customers (as in a competitive market) are held no lower than the full additional costs of taking on that additional business, and prices for inelastic demand customers (as in a monopoly market) are held no higher than average total cost, then there can be no internal subsidization. 29 The LRIC pricing method is designed precisely to determine the full additional costs of additional business in a competitive market. If properly applied, it is sufficient to prevent cross-subsidies.
The LRIC method is a specific application, in modified form, of marginal cost pricing. Samuelson describes the rationale, from an individual firm’s perspective, for making pricing decisions according to extra or marginal costs, while viewing sunk costs — such as the common and joint costs of AT&T — as “bygones”:
Let bygones be bygones. Don’t look backward. Don’t moan about your sunk costs. Look forward. Make a hard-headed calculation of the extra costs you’ll incur by any decision and weigh these against its extra advantages. Cancel out all the good things and bad things that will go on anyway, whether you make an affirmative or negative decision on the point under consideration. 30
The LRIC method is designed to perform exactly this sort of calculation, to compare additional revenues — i. e., rates for competitive services — with the incremental costs of providing those services. The FDC method, by contrast, would force AT&T to focus on the “bygones” while its competitors are free to employ a more rational form of competitive pricing.
Kahn concludes that long run marginal costs are the proper starting point for assessing the validity of competitive rate reductions by regulated companies in competitive markets. 31 Society has an interest in communications services being provided at lowest possible cost, and economic efficiency considerations require “that no business be turned away that covers the costs to society of providing that service.” 32 LRIC is calculated to achieve precisely this efficiency, FDC is not. It is important to note that Kahn does envision limited circumstances in which long run marginal costing would not be appropriate, 33 but the Commission did not even consider whether such special circumstances might make this case an exception to the general desirability of marginal or incremental cost pricing.
The Commission did not directly attack the desirability of marginal cost pricing. At some points in its decision, the Commission appeared to concede the theoretical merits of marginal cost pricing, 34 while at other points it implied that marginal cost pricing would necessarily lead to cross-subsidies. Rather than attack the marginal cost pricing concept directly, the Commission objected that LRIC is not truly a marginal cost technique, 35 and it is true that LRIC does not price according to the cost of producing the marginal single unit of service. But in fact LRIC, by focusing on long run increments to joint and common costs as well as the additional variable costs of producing the extra units, sets a cost figure that is likely to be higher than pure marginal costs. 36 Thus LRIC leaves even less *275 chance of cross-subsidies and predatory pricing by requiring higher rates than pure marginal cost. This factor explains why it is that those with an institutional concern for the prevention of predatory pricing, such as the Antitrust Division of the Justice Department, can support LRIC pricing in the present case.
In short, the FCC’s assumption that FDC pricing rather than LRIC is necessary to prevent cross-subsidies is clearly contrary to a common sense evaluation of the facts and to commonly accepted economic theory. As a result the Commission has not shown its standard for minimum rates to be rationally related to the prevention of cross-subsidization. By choosing a standard that requires high rates without a showing of reasonable connection to a legitimate regulatory purpose, the Commission has unjustifiably restrained competition and imposed an unjustifiable price burden on AT&T’s competitive service customers. (Hence the position and role of the Antitrust Division in this case.) The Commission’s invalid assumption and conclusions regarding FDC and LRIC pricing reflect a thoroughly uncertain grasp of the fundamentals of pricing in the industry it regulates. By accepting the Commission’s decision at face value, the majority does nothing to shake it from these fallacies.
The Commission raised several more objections against LRIC pricing in addition to its assumption that LRIC pricing would allow cross-subsidies. While it might be permissible for the Commission to choose FDC pricing after deciding that these other factors outweigh the positive value of LRIC, a rational balancing is not possible unless one first acknowledges the positive value of LRIC — i.e., that it prevents cross-subsidies without impairing competition as severely as does FDC.
For purposes of clarification, it is appropriate to comment on the merit of these additional arguments against LRIC. First is the argument that raising prices in the competitive market will not lose competitive service customers for AT&T. The FCC considers it “arbitrary” of AT&T to single out competitive service customers for marginal cost pricing; why, the Commission asks, doesn’t the firm use marginal cost pricing for monopoly customers as well? 37
The simple answer is that it is only the competitive service customers who are likely to be lost through slight increases in prices. The ground for difference in treatment is that elasticity of demand is greater in competitive markets than in monopoly markets. 38 It is only in the competitive market that AT&T must respond to competition by keeping rates at a low competitive level. Thus it is simply the facts of economic life, and not an invidious bias against monopoly service customers, that induces AT&T to lower its competitive service prices toward the LRIC level. As we have seen pricing anywhere above LRIC cannot result in a subsidy. The Commission, on the other hand, appears to reject elasticity-based pricing generally, when it states (quoting the Recommended Decision) that
[wjhen economists suggest (as in this proceeding) the use of ‘inverse elasticity rules’ for price discrimination among services or classes of customers to allow the company to meet its overall revenue requirement, they are suggesting that those customers with inelastic demand subsidize those with elastic demand. 39
To follow the reasoning of this statement to its logical conclusion would prohibit elasticity-based pricing wherever found for all regulated firms — such a direct contradiction of the normal workings of competitive pricing is not justified by the record or reasoning of the Commission in this case.
This line of reasoning leads to a further grave flaw in the Commission’s opinion. By requiring an increase in competitive service rates to the FDC level in order to prevent cross-subsidies, the Commission im *276 plicitly assumes that this increase in rates will produce a rise in revenues. But this will happen only if demand is inelastic. Not only has the Commission shown no reason to believe that demand is inelastic, but it admits that this is a competitive market, which by nature should tend toward elastic demand. Thus the Commission has offered no evidence to show a probability or even a reasonable hope that raising AT&T’s competitive service rates to the FDC level will not price AT&T out of the market, and thus eliminate any and all contribution that the competitive service customers were making to common and joint costs.
The Commission perhaps recognizes this flaw in its reasoning, for it argues that even if AT&T loses competitive service customers with price increases in the competitive market, its expanding monopoly services will soon utilize the idle facilities. 40 But the existence of idle facilities is not the important factor here; the important fact is that without competitive service customers, the firm loses the contribution to common or joint costs that is provided whenever competitive service customers pay a rate even slightly higher than LRIC. Thus the loss of competitive service customers will hurt monopoly service customers, whether or not the competitive service facilities remain idle only temporarily.
Finally, the FCC maintains that whatever the merits of LRIC pricing, the Commission’s other objections justified the use of FDC pricing. 41 The Commission emphasizes the objective of accountability, which requires that any costing methodology be verifiable so that the FCC can ensure that AT&T will not allow cross-subsidies to arise in its pricing structure. 42 The Commission alleges that LRIC is subjective, in the sense that it “is heavily dependent upon judgmental decisions which cannot be verified or tested by the Commission.” 43 While such considerations may certainly enter into a rational weighing of the merits of LRIC versus FDC, the Commission should not overlook the subjectivity and arbitrariness that are unavoidable in the FDC method, which performs the inherently arbitrary task of assigning unattributable costs to particular services. And if current reporting procedures do not generate data for the verification of LRIC pricing, the FCC should consider the possibility of developing procedures that can generate such data.
The FCC also maintains that its allowance for price reductions below FDC upon a showing of competitive necessity supports the rationality of its decision. 44 There are two criteria for allowing a competitive necessity exception to the general requirement of FDC pricing: (1) an alternative supply source that customers will utilize unless they are given the benefits of price discrimination, and (2) a benefit from discrimination for users who are discriminated against. 45 The problem is that the Commission’s requirement of a showing of a “real competitive threat” 46 is very difficult to satisfy, and was not satisfied in this case, even though the market in question was admittedly competitive. The FCC’s standards ignore that if there is a legitimate concern on the part of the firm for pricing at low levels, it must be because there are actual or potential competitors. The FCC conceded that the private line services market is a competitive market when it initiated Docket No. 16,258 47 But it apparently will allow competitive pricing in competitive markets only upon a showing of exceptional circumstances. The “zone of reasonableness,” that the Commission claims to *277 allow for prices, 48 thus turns out to exist only in exceptional cases.
CONCLUSION
As stated above, it is possible that a rational balancing of all factors might favor FDC pricing, despite the economic pro-competition advantages of LRIC and its ability to prevent cross-subsidies. The economic muddle of the FCC’s analysis in this case, however, makes it clear that the Commission never recognized the economic merits of LRIC, and that it assumed FDC pricing to be necessary to prevent cross-subsidies. It thus was unable to give rational consideration to the balancing of the needs to promote competition and to prevent cross-subsidies, which under all standards of judicial review makes it impossible for this court to affirm agency action demonstrably resting on no rational base.
The primary effect of forcing AT&T’s competitive service prices up to the FDC level is to protect its competitors in that market. In the words of the Department of Justice, “the Commission confuses protecting competitors and protecting competition.” 49 This result contrasts sharply with the Commission’s earlier announced policy of refusal to provide a “protective umbrella” for new entrants into AT&T’s markets. 50 And it imposes a high cost on AT&T’s consumers, a cost not justified on the record before us.
The complexity of analysis in this case might lead one to question the certainty with which a court can say that the Commission acted arbitrarily. 50a But in this case the arbitrariness of the FCC’s reasoning is very simply and clearly demonstrated by the absurdity of its result: under the FCC’s prescribed costing procedures, the maximum level of rates set to prevent monopoly pricing is exactly identical to the minimum level established to prevent pred- ■ atory pricing and cross-subsidies. The Justice Department accurately summarizes the effect of this ruling as follows: “AT&T, subject to extremely narrow waiver provisions, will as a practical matter have virtually no freedom to price a service any differently than it prices its basic monopoly service, notwithstanding possible efficiencies,” 51 If AT&T raises rates above the prescribed level, it is guilty of monopoly pricing; if it lowers them below this level, it is guilty of cross-subsidizing and, in effect, predatory pricing — or so decrees the Commission.
Even apart from economic analysis, common sense alone suffices to show that there must be some range of prices, within which firms have discretion to choose in a competitive market without being immediately accused of either monopoly pricing or predatory pricing. Otherwise business judgment and price competition have no role whatever in the market. Much as courts should normally defer to an agency’s expertise, there is some limit beyond which a court must step in. When an agency arrives at a decision so contrary to common sense and economic logic as this, we have reached that limit. 52
*279 I would therefore vacate the relevant parts of the Commission’s decision and remand; hence I respectfully dissent from Part II A and concur separately in Part II C of the majority opinion.
. 61 F.C.C.2d 587 (1976).
. See Permian Basin Area Rate Cases, 390 U.S. 747, 776 , 88 S.Ct. 1344, 1364 , 20 L.Ed.2d 312 (1968), cited in Maj. Op. at 1228.
.Id. (citations omitted).
. See Brief for Petitioner AT&T at 13, 44 — 45; Reply Brief for Petitioner AT&T at 16-23.
. See Brief for Respondent United States of America at 32 — 40.
. 47 U.S.C. § 201 (b) (1976).
. Id. § 202(a).
. See, e. g., I A. KAHN, THE ECONOMICS OF REGULATION 20 (1970); In re American Tel. & Tel. Co., 61 F.C.C.2d at 77 (quoting Recommended Decision).
. See, e. g., F. SCHERER, INDUSTRIAL MARKET STRUCTURE AND ECONOMIC PERFORMANCE 273-74 (1970).
. Brief for FCC at 48.
. See In re American Tel. & Tel. Co., 61 F.C. C.2d at 612.
. See id. at 616.
. See id. at 618-19 (quoting Recommended Decision).
. See id. at 607.
. See id. at 619 (quoting Recommended Decision).
. See I A. KAHN, supra note 8, at 77-79.
. See In re American Tel. & Tel. Co., 61 F.C. C.2d at 619-26.
. Id. at 624.
. See id. at 652.
. Id.
. Id. at 669.
. Brief for FCC at 53.
. “. . .it was a matter for the Commission, in its discretion, to determine whether this aspect of LRIC pricing in mixed economic environments poses a serious drawback.” Maj. op. at 1229 n.13. Contrary to the majority’s attempted evasion of hard economic analysis, the FCC’s assumption, as evidenced in the quoted passages above and throughout the FCC opinion, is the key to the economic errors on which the decision is founded.
. See In re American Tel. & Tel. Co., 61 F.C. C.2d at 610.
. See generally P. SAMUELSON, ECONOMICS 381-83 (1976).
. I A. KAHN, supra note 8, at 159 (emphasis in original).
. Id. at 162.
. See generally id. at 63-86. See also In re American Tel. & Tel. Co., 61 F.C.C.2d at 670-71 (Hooks, Comm’r, dissenting) (citing authorities).
. Alchian says: “The general rule is simple: To minimize the costs of any specified amount of X, produce it with the lowest marginal costs. . . In brief, producers should produce amounts that equate their marginal costs (or comes as near as possible to equating them.). That will result in no waste — that is, in minimum cost.” A. ALCHIAN, EXCHANGE AND PRODUCTION 172-77 (1977).
Samuelson says: “Our analysis has demonstrated the following remarkable truth: Only when prices of goods are equal to Marginal Costs is the economy squeezing from its scarce resources and limited technical knowledge the maximum of outputs. Only when each source of industry output has had its rising MC [marginal cost] equated to any other source’s MC— as will be the case when each MC has been set equal to the common P [price] — can the industry be producing its total Q [quantity] at minimum Total Cost. Only then will society be out on its production-possibility frontier and not efficiently inside the frontier. This equal-marginal-cost dictum is as applicable to a communist, socialist, or fascist society as to a capitalistic society.” P. SAMUELSON, supra note 24, at 462 (emphasis in original)..
.“These are the rules that most authorities have adopted, and with which we concur. One implies the other; as long as the favored customers pay their full additional costs, the others cannot on this account be led to pay more than the costs of serving them in the absence of discrimination. And both together imply the condition that discrimination be permitted only as long as it imposes no burden on the customers being discriminated against. Such a rule would prohibit internal subsidization.” I A. KAHN, supra note 8, at 142-43 (emphasis in original). See also id. at 142 n.37, and sources cited therein.
. P. SAMUELSON, supra note 24, at 498 (emphasis in original).
. See I A. KAHN, supra note 8, at 159-75.
. Id. at 160-61.
. See id. at 166-70.
. See, e. g., In re American Tel. & Tel Co., 61 F.C.C.2d at 636 n.80.
. See id. at 621-24.
. See I A. KAHN, supra note 8, at 75, 143-44.
. See In re American Tel. & Tel. Co., 61 F.C. C.2d at 614, 626-27, 629, 639.
. See I A. KAHN, supra note 8, at 140-44.
. See In re American Tel & Tel. Co., 61 F.C. C.2d at 642.
. See id. at 640.
. See id. at 636 n.80.
. See Brief for FCC at 30, 50, 54; In re American Tel. & Tel. Co., 61 F.C.C.2d at 610-12.
. In re American Tel. & Tel. Co., 61 F.C.C.2d at 632.
. See Brief for FCC at 53.
. See In re American Tel. & Tel. Co., 61 F.C. C.2d at 590.
. Id. (emphasis in original).
.See id. at 592.
. Id at 641.
. Reply Brief for Respondent United States of America at 9 n.10.
. See Specialized Common Carrier Services, 29 F.C.C.2d 870 , 915 (1971).
. For example, in American Commercial Lines, Inc. v. Louisville & N.R.R., 392 U.S. 571 , 88 S.Ct. 2105 , 20 L.Ed.2d 1289 (1968), the Supreme Court stated in dictum that it was reluctant to pass on the merits of marginal cost economic arguments. Id. at 586 n.16, 88 S.Ct. at2113n.l6. But the Court added at the end of this footnote: “Our discussion here should not be interpreted as a rejection of the basic economic points made by the railroads. It is merely intended to illustrate the desirability of having the initial resolution of these issues made by a tribunal, and in a proceeding, more suitable than the present one.” Id. at 589 n.16, 88 S.Ct. at 2114 n.16 (emphasis added).
. Brief for Respondent United States of America at 35 (emphasis in original) (citations deleted). See also In re American Tel. & Tel. Co., 64 F.C.C.2d at 978 & nn. 13-14 (1977) (on petitions for reconsideration).
. A simple example, in a situation familiar to judges, lawyers, law teachers, and even law students, will make clear the above economic exposition, even if the reader has no background in economics whatsoever. This exam- *278 pie will illustrate the fundamental difference between these two cost and pricing concepts, and show the basic fairness and ultimate economic utility of long-range incremental cost pricing.
Suppose that Judge X is called by the moot court board at Anywhere Law School and invited to come to Anywhere and sit as a judge in the moot court finals. The board representative states that all of Judge X’s actual out-of-pocket expenses will be paid. Judge X, interested in giving a pleasant outing to his wife, inquires if the invitation, all expenses paid, includes Mrs. X in the hospitality. The person in charge replies, “I am sure that Mrs. X would be very welcome, and we’ll be glad to make any arrangements for her at the hotel, but at the moment I can’t say whether the moot court board is in a position to pick up her expenses at the hotel.” Judge X replies, “I would like to bring Mrs. X anyway, so please make the reservations for her and we’ll see later where the expense is to be charged.”
Judge and Mrs. X attend the moot court, the bill at the hotel for the two is $150 a day, full American plán, and Judge X carefully notes that it would have been $125 a day for him alone. On return he sends in his itemized statement of expenses and notes that the question of whether the moot court board is picking up the expenses of Mrs. X had not been resolved — if they are, he should be reimbursed at $150 a day; if not, he should be reimbursed at $125 a day.
The moot court board replies, saying that they are very sorry but they cannot pick up the hotel expenses of Mrs. X and that they are enclosing a reimbursement for Judge X’s expenses calculated on a basis of $75 a day. Judge X doesn’t understand this. If he had gone alone, it is very clear that he would have been reimbursed at $125 a day for himself alone, because this is what the charge for the room, full American plan, plainly was. The added increment for Mrs. X was only $25 a day at the most. It is clear that the moot court board would have had to pay $125 a day for Judge X alone, and this was their original invitation on this basis.
The moot court board replies, “Judge X, you don’t understand fully distributed costs. It’s obvious that the facilities at the hotel, including room and meals and all services, were enjoyed equally by you and Mrs. X. One half of $150 a day is $75 a day, which is the cost properly allocated to you alone. It is true that if you had come by yourself, the cost would have been $125 a day, and we would have reimbursed you on that basis, but you didn’t come alone, you added another pleasurable increment to your visit here, namely, the company of Mrs. X, and that part is properly chargeable to Mrs. X and you, and not to the moot court board. We believe in the economics of fully distributed costs, and fully distributed costs means $75 allocated to Mrs. X and $75 allocated to Judge X. We are paying only for Judge X.
Judge X further protests, “If I had known you were going to charge on this basis, I would never have brought Mrs. X at $75 a day. I would have come by myself at the agreed all-expense rate, payable at $125 a day for me, and never brought Mrs. X into the picture. If there is anything I’m clear about, it’s the unfairness of this retroactive application of your theory on costs. You did say it was undecided whether you could pay for Mrs. X or not, but you never said that you were going to charge for my expenses on a different basis if I brought Mrs. X. This retroactive application is the most unfair thing of all.”
The moot court board replies, “If there is anything clear from our point of view, it is that you were on full notice that we might or not pay Mrs. X’s expenses, and if we couldn’t pay Mrs. X’s expenses, surely you must realize that we were at liberty to use the logical method of fully distributed costs in calculating the expense properly chargeable to you and to her. The fact that we made up our minds later as to whose expenses to pay and by which method to calculate the costs is immaterial. You were on full notice of this all along.”
The parallel with the AT&T situation is obvious. The cost for monopoly services is fixed, and will be the same whether the competitive services are added by AT&T or not. The cost for the monopoly services is the equivalent of the $125 which would have been the hotel charge for Judge X; the cost for the competitive services would be the additional $25 which would be added if Mrs. X enjoyed the use of the same services. From the hotel’s point of view, the additional cost for Mrs. X’s presence is only the additional linens and the food, and therefore the incremental cost for her presence is a relatively small amount compared to the basic cost of providing that one room and facilities for one person. (The hotel might calculate that the presence of Mrs. X would generate sales in the shops on the hotel premises and thus add a small amount to hotel revenues, and thus the hotel could encourage her presence by charging even less than the cost of the linens and food and still make a profit on the incremental services. Similar comparisons might be made to AT&T services.)
To apply fully distributed costs to the stay of the couple at the hotel is economic nonsense; it is unquestionably true that considered ab initio the cost of providing the room and food for the two persons can be divided equally, $75 apiece, but this bears no relation to the economic logic of the way to conduct a hotel business or to conduct a moot court board’s business either. What the moot court board was faced with from the start was paying the total expenses of Judge X, which amounted to $125 at the price charged by the hotel. Similarly, the price charged by the hotel for Judge X individually was what it cost to put one person in the room *279 and provide meals and all services, with a reasonable profit.
Both the hotel and the moot court board should logically and sensibly run their business on an incremental cost basis, just as is advocated by AT&T and the Antitrust Division in our case. The basic cost of having Judge X come to the moot court is going to be $125 a day whether Mrs. X comes or not. The board is logically forced to pay this price. The advent of Mrs. X is something entirely within the control of Judge X; she can come or not, and if she does come, there is no moral or economic right of the moot court board to profit by the incremental service the hotel is providing for Mrs. X by reducing the cost allocated to Judge X’s presence, which logically still remains at $125 a day and is not either economically or equitably reducible to $75.
Similarly, the charges which the monopoly customers of AT&T have been paying and which have been previously determined as fair, based on the costs of AT&T in providing these services, should not necessarily be reduced simply because AT&T can inaugurate other services in the competitive market. The costs properly allocated to the new competitive services of AT&T are incremental costs, not fully distributed costs, because the costs of the monopoly services should remain the same irrespective of whether AT&T enters competitive markets or not. (Actually, as shown by the detailed economic exposition above, there is hope that the competitive services of AT&T, with incremental cost pricing, will definitely contribute to a lowering of costs for the monopoly service customers.)
Of course, as Judge X would probably indignantly point out in our hypothetical case, the retroactive determination by the moot court board to reimburse him on the basis of fully distributed costs, i. e., only $75 for his half of the hotel room and services, is the most unfair and illogical action of all. While Judge X did understand that the moot court board might or might not pay for Mrs. X, he doubtless thought that surely the question of the cost of his own visit was fully settled, and would not be affected by whether the incremental cost of Mrs. X was added. This compares with AT&T knowing that the whole question of an ultimate costing was yet to be decided, but surely AT&T should not be penalized retroactively on a fully distributed cost theory, when the cost of its monopoly services would have remained the same over the years, whether AT&T had added the incremental services or not.