Northeastern Telephone Company v. American Telephone and Telegraph Company

IRVING R. KAUFMAN, Circuit Judge: In Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979), cert. denied, 444 U.S. 1093, 100 S.Ct. 1061, 62 L.Ed.2d 783 (1980), this Court plumbed the crosscurrents of Section 2 of the Sherman Act, 15 U.S.C. § 2, holding that dominant firms, having lawfully acquired monopoly power, must be allowed to engage in the rough and tumble of competition. This case presents us with the opportunity to elucidate and to apply the rationale of Berkey in the context of the American telecommunications industry. It presents an antitrust suit brought by Northeastern Telephone Co., a relatively small supplier of telephone equipment, against a mammoth and legendary enterprise — the American Telephone & Telegraph Co. (AT&T). Joined as defendants were Western Electric, the manufacturing arm of the Bell System, and Southern New England Telephone Co. (SNET), the local Bell affiliate serving virtually all of Connecticut. For much of their existence, these entities have escaped the rigors of competition, sheltered in part by tariffs filed with the Federal Communications Commission (FCC). In 1968, however, the Commission exposed a significant portion of their business to competitive pressures. We are here called upon to apply the teachings of Ber-key to the actions appellants took in response to this infusion of rivalry. To oversimplify somewhat, the question presented is whether appellants did no more than to engage in vigorous competition, or instead, attempted to subvert the competitive process by unfair or unreasonable means. Because we hold that Northeastern has, for the most part, failed to prove that appellants’ conduct exceeded the bounds of competitive propriety outlined in Berkey, we reverse the judgment below in several major. respects. As to the design of the “protective coupler arrangement” required by appellants for use with Northeastern’s telephone equipment, we remand for a new trial on liability and damages. I. FACTUAL BACKGROUND AND PROCEEDINGS BELOW A brief overview of the history of competition in the relevant market is necessary to our. analysis of the legal questions raised in this appeal. For many years, customers of Bell System affiliates, 1 including customers of SNET, 2 were prohibited from connecting their own terminal equipment into the Bell communications network, even if those items were identical to the equipment supplied by local Bell affiliates. This prohibition, which was embodied in tariffs filed with the FCC, completely foreclosed competition in the terminal equipment market. In 1965, however, at the urging of a federal district court in the Northern District of Texas, the FCC began an investigation into the reasonableness of this practice. Three *80 years later, the Commission invalidated the tariffs, ruling that AT&T’s refusal to allow interconnection of customer-provided equipment violated § 201(b) of the Federal Communications Act, 47 U.S.C. § 201(b). Thus was the “interconnect industry” born. See In re Use of Carterfone Device, 13 F.C.C.2d 420, reconsideration denied, 14 F.C.C.2d 571 (1968). Many companies were attracted to the new market. Some, like Nippon Electronics and International Telephone & Telegraph (ITT) were huge conglomerates; others, like Northeastern Telephone, were diminutive firms. Indeed, Northeastern was among the smallest of the new entrants. It was formed by two Connecticut businessmen in 1972 on a total capital investment of $1,000. Its first corporate headquarters were in the basement of an old church, and its founders primarily did maintenance work on equipment sold by ITT. Its revenues that first year were only $70,000. In succeeding years, however, the company expanded dramatically, adding approximately one new office per annum. It is now headquartered in Milford, Connecticut, and its revenue in its seventh year of operation, 1978, exceeded $3,000,000. The competitive battle between Northeastern and SNET was waged on two fronts; public branch exchanges (PBXs) and key telephones. 3 A PBX consists of three elements: the equipment located on a customer’s premises and used to switch calls from one telephone line to another; a switchboard which controls that operation; and the associated telephone sets and wiring. Key telephones, which are a familiar sight in business offices, are equipped with keys or buttons to give the set access to more than one telephone line. During the first few years of their rivalry, both Northeastern and SNET obtained PBXs from outside suppliers — often Nippon Electronics. But in January 1977, SNET submitted, and the Connecticut Division of Public Utilities Control (DPUC) approved, a tariff enabling the utility to offer two PBXs manufactured by Western Electric— the Dimension 100 and the Dimension 400. 4 These units, particularly the Dimension 400, were well-received by Connecticut businessmen. We are told that SNET marketed over one hundred Dimension PBXs between March 1977 and February 1978. The competitive environment in the business terminal equipment market, while not fatal to Northeastern’s continued survival, was doubtlessly hostile. Metaphorically, Northeastern was a mosquito challenging an elephant. Even in its best year, its annual revenues from all of its operations were less than one-twentieth of the returns SNET earned in the terminal equipment market alone. But similar size disadvantages face any aspiring entrant wishing to dislodge a dominant firm. The antitrust laws assume these risks, and Northeastern must be taken to have accepted them. Ap-pellee alleges, however, it was also the victim of business practices not countenanced by the Sherman Act. Specifically, Northeastern contends that SNET’s prices for its Dimension PBXs and its key telephones were predatorily low, and that the customized payment option SNET offered to some of its business customers, the so-called “two-tier payment plan,” had anticompeti-tive effects. Northeastern also complains that other aspects of appellants’ activities were intended to stifle competition: their advertising and marketing methods, their introduction of new products, and their use of their monopoly power over telephone service to distort competition in the business equipment market. Finally, Northeastern challenges conduct related to the revised tariffs AT&T filed in response to the FCC’s Carterfone decision. After the Commission had invalidated the *81 proscription against interconnection of customer-owned terminal equipment in 1968, AT&T proposed tariffs requiring that all such equipment be interconnected via a “protective coupler arrangement,” to be provided, installed, and maintained by the local operating companies at the customer’s expense. These couplers were intended to protect the telecommunications network from certain electrical problems that might result from the use of faulty or incompatible equipment. Although several telecommunications firms objected to these tariffs, the FCC allowed them to take effect pending the outcome of a study it had commissioned by the National Academy of Sciences (NAS). The Commission noted, however, that in taking this approach it was “not giving any specific approval to the revised tariffs.” In re A.T.&T. “Foreign Attachment” Tariff Revisions, 15 F.C.C.2d 605, 610 (1968), reconsideration denied, 18 F.C.C.2d 871 (1969). The NAS consumed the next four years, from 1968 to 1972, in preparing its report. It concluded, finally, that although the protective coupler requirement was “an acceptable way of assuring network protection,” a registration system also merited consideration. Under such a system, the FCC would promulgate minimum specifications designed to protect the telecommunications network from electrical harm. Equipment meeting these standards would be exempt from the coupler requirement, but couplers would still be mandatory for unregistered items. Three years after adopting the NAS report, the FCC implemented this recommendation and invalidated AT&T’s post- Carterfone tariffs. See In re Proposals for New or Revised Classes of Interstate and Foreign Message Toll Telephone Service and Wide Area Telephone Service, First Report and Order, 56 F.C.C.2d 593 (1975); Second Report and Order, 58 F.C.C.2d 736, on reconsideration, 61 F.C.C.2d 396 (1976), 64 F.C.C.2d 1058, aff’d sub nom. North Carolina Utilities Commission v. F.C.C., 552 F.2d 1036 (4th Cir.), cert. denied, 434 U.S. 874, 98 S.Ct. 222, 54 L.Ed.2d 154 (1977). Northeastern does not challenge the legality of the invalidated tariff. It contends instead that the protective couplers were intentionally overdesigned, making them unnecessarily expensive and subject to break down. In particular, Northeastern complains that the couplers required an external power source 5 and that they had six-wire leads while Northeastern’s PBXs were designed for two-wire interconnection. These features, Northeastern maintains, unreasonably hampered its efforts to compete in the business terminal equipment market. Northeastern instituted the present suit in 1975, alleging that SNET, in violation of Section 2 of the Sherman Act, 15 U.S.C. § 2, had monopolized and had attempted to monopolize the business terminal equipment market in Connecticut. 6 Appellee charged further that SNET, AT&T, and Western Electric had conspired to monopolize that market, also in violation of § 2, and had conspired to restrain trade, in contravention of § 1, 15 U.S.C. § 1. The case was originally assigned to Judge Daly, before whom appellants moved to dismiss the complaint on the ground that their allegedly anticompetitive conduct was exempt from antitrust scrutiny. Judge Daly denied that motion in November 1978, D.C., 477 F.Supp. 251, and, after ten more months of preliminary maneuvering, the trial commenced before Judge Eginton and a jury. Two weeks later, Northeastern filed an amended complaint, adding its objections to appellants’ Dimension PBX tariffs. By agreement of the parties, the trial was bifurcated. Northeastern’s proof at the liability phase centered on the six types of conduct it alleged to be anticompetitive: pricing, advertising, marketing, introduction of new products, SNET’s alleged use of its utility *82 function to impede competition, and appellants’ design of the protective coupler. The presentation of this evidence and of appellants’ proof in opposition required almost three weeks. When both sides finally rested, the jury was asked to respond to fourteen interrogatories. It returned a verdict for Northeastern on all four of the claims, finding that each of appellants’ six challenged activities was anticompetitive. Upon presentation of further evidence on damages, the jury awarded appellee $5,515,-692. Of this amount, $3,368,906 was for lost profits; $2,146,786 was for damage to Northeastern’s “going concern value.” Pursuant to Section 4 of the Clayton Act, 15 U.S.C. § 15, the district court ordered that the amount be trebled, to the sum of $16,547,076. Appellants then moved for judgment n. o. v. or for a new trial, renewing their claims of antitrust immunity and arguing that Northeastern had not presented sufficient evidence to support the verdict. Judge Eg-inton denied the motions, D.C., 497 F.Supp. 230, holding that except with respect to the evidence concerning SNET’s pricing of key telephones, the jury had an ample basis upon which to predicate its findings of liability and damages. He then awarded Northeastern an additional $747,813.38 in attorneys’ fees, as authorized by 15 U.S.C. § 15. This appeal followed. II. IMPLIED ANTITRUST IMMUNITY BY VIRTUE OF FEDERAL REGULATION A threshold question prevents our proceeding directly to Northeastern’s Sherman Act claims. Appellants assert that the design of the protective coupler cannot be attacked on antitrust grounds because implementation of the coupler requirement was subject to review by the Federal Communications Commission. 7 It is to this issue that we now turn. Appellees claim no explicit exemption; it is on a more elusive defense that they base their parry: “implied immunity.” This principle represents an effort to resolve the inherent conflict between the Sherman Act’s mandate of robust competition and the “public interest” standard underlying governmental regulation of business activity. The touchstone of this analysis is Congressional intent, see Otter Tail Power Co. v. United States, 410 U.S. 366, 93 S.Ct. 1022, 35 L.Ed.2d 359 (1973), since the principle is founded on the notion that, in some circumstances, a Congressional delegation of regulatory authority carries with it the implication that the antitrust laws shall not apply to the conduct thus regulated. There are two narrowly-defined situations in which “repeal” of the Sherman Act will be inferred: first, when an agency, acting pursuant to a specific Congressional directive, actively regulates the particular conduct challenged, see, e. g., Gordon v. New York Stock Exchange, Inc., 422 U.S. 659, 685-86, 688-89, 95 S.Ct. 2598, 2614, 45 L.Ed.2d 463 (1975), and second, when the regulatory scheme is so pervasive that Congress must be assumed to have forsworn the paradigm of competition, see, e. g., United States v. National Association of Securities Dealers, Inc., 422 U.S. 694, 730, 95 S.Ct. 2427, 2448, 45 L.Ed.2d 486 (1975); Otter Tail Power Co. v. United States, supra, 410 U.S. at 373-74, 93 S.Ct. at 1027-1028. In either case, immunity will not lightly be inferred. “Repeal of the antitrust laws by implication is not favored and not casually to be allowed. *83 Only when there is a ‘plain repugnancy between the antitrust and regulatory provisions’ will repeal be implied.” Gordon v. New York Stock Exchange, Inc., supra, 422 U.S. at 682, 95 S.Ct. at 2611, quoting United States v. Philadelphia National Bank, 374 U.S. 321, 350-51, 83 S.Ct. 1715, 1734, 10 L.Ed.2d 915 (1963). As a further limitation, “[r]epeal is to be regarded as implied only if necessary to make the [regulatory scheme] work, and even then only to the minimum extent necessary.” Silver v. New York Stock Exchange, 373 U.S. 341, 357, 83 S.Ct. 1246, 1257, 10 L.Ed.2d 389 (1963). Beyond setting out these basic principles, cases involving federal regulatory bodies other than the FCC are of limited usefulness in the present inquiry. See Essential Communications Systems, Inc. v. American Telephone & Telegraph Co., 610 F.2d 1114, 1116-17 (3d Cir. 1979); L. Sullivan, Antitrust 743-44 (1977). Whenever claims of implied immunity are raised, they must be evaluated in terms of the particular regulatory provision involved, its legislative history, and the administrative authority exercised pursuant to it. See Gordon v. New York Stock Exchange, Inc., supra; United States v. Philadelphia National Bank, supra. This analysis has been undertaken with regard to the communications industry in several recent and comprehensive opinions, which trace Congressional efforts to regulate this area from their origin in the MannElkins Act of 1910, Pub.L.No.218, 36 Stat. 539, through the Willis-Graham Act of 1921, Pub.L.No.15, 42 Stat. 27, to their present incarnation — the Federal Communications Act of 1934, 47 U.S.C. §§ 151-609 (the 1934 Act). See, e. g., Essential Communications Systems, Inc. v. American Telephone & Telegraph Co., supra, 610 F.2d at 1116-21; Litton Systems, Inc. v. American Telephone & Telegraph Co., 487 F.Supp. 942, 947-50 (S.D.N.Y.1980). See also MCI Communications Corp. v. American Telephone & Telegraph Co., 462 F.Supp. 1072 (N.D.Ill.), aff’d sub nom. American Telephone & Telegraph Co. v. Grady, 594 F.2d 594 (7th Cir. 1978) (per curiam), cert. denied, 440 U.S. 971, 99 S.Ct. 1533, 59 L.Ed.2d 787 (1979); United States v. American Telephone & Telegraph Co., 461 F.Supp. 1314 (D.D.C.1978). No purpose would be served by repeating this exegesis here. It is sufficient to state that having conducted an independent review of the 1934 Act and its legislative history, we find that immunity cannot be inferred on either of the two grounds previously discussed. We note first that the Act does not expressly authorize the FCC to approve protective coupler designs that unreasonably restrict competition. Compare Gordon v. New York Stock Exchange, Inc., supra (section 19(b)(9) of Securities Exchange Act of 1934 directed Securities Exchange Commission to supervise specific conduct challenged; if not immunized, conduct would have constituted per se violation of Sherman Act). While this observation is unsurprising, since protective couplers were unknown in 1934, it does rebut appellants’ argument that immunity may be inferred on the basis of specific Congressional authorization. Appellants’ second assertion, that they are entitled to immunity on the vague ground of pervasive regulation, is also without merit. The pervasiveness of a regulatory scheme is not susceptible of precise quantification, see 1 P. Areeda & D. Turner, Antitrust Law 1224 (1978). But it is not meant to be, since regulation is not an end in itself; it is only a means of inferring that Congress intended to free the regulated industry from the discipline of competition. The more focused inquiry is whether the industry is so extensively regulated that application of the antitrust laws would be incompatible with the regulatory framework, that is, whether immunity must be inferred to make the system work. See United States v. National Association of Securities Dealers, Inc., supra, 422 U.S. at 734, 95 S.Ct. at 2450; Silver v. New York Stock Exchange, supra, 373 U.S. at 357, 83 S.Ct. at 1257. A brief review of the FCC’s treatment of the protective coupler tariff demonstrates that applying the Sherman Act to the matter of coupler design would not frustrate federal regulation of the telecommunications industry. AT&T filed this tariff in *84 response to the FCC’s decision barring telecommunications carriers from prohibiting the interconnection of customer-owned terminal equipment. See Carterfone, supra. The Commission, realizing that an adequate study of the revised tariff would take several years, 8 allowed it to take effect pending further investigation. It emphasized, however, that “in doing so, we are not giving any specific approval to the revised tariffs.” AT&T “Foreign Attachment” Tariff Revisions, supra, 15 F.C.C.2d at 610-11. Seven years later, the FCC invalidated the tariff and substituted a registration system which freed equipment meeting certain technical specifications from the protective coupler requirement. That the Commission never approved the protective coupler tariff demonstrates that no conflict will arise between the Federal Communications Act and the antitrust laws if we hold that appellants are subject to antitrust liability for designing the coupler as they did. See Essential Communications Systems, Inc. v. American Telephone & Telegraph Co., supra, 610 F.2d at 1124. See also Litton Systems, Inc. v. American Telephone & Telegraph Co., supra. Accordingly, we reject appellants’ claim that the design of the protective coupler escapes scrutiny under the Sherman Act. 9 III. NORTHEASTERN’S SHERMAN ACT CLAIMS A. Framework for Analysis Having demonstrated that appellant’s conduct is not beyond the reach of the antitrust laws, we turn to an examination of the proof at trial. In doing so, of course, we must “view the evidence in the light most favorable to [Northeastern] and . .. give it the benefit of all inferences which the evidence fairly supports, even though contrary inferences might reasonably be drawn.” Continental Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 696, 82 S.Ct. 1404, 1409, 8 L.Ed.2d 777 (1962). Accord Michelman v. Clark-Schwebel Fiber Glass Corp., 534 F.2d 1036, 1042 (2d Cir.), cert. denied, 429 U.S. 885, 97 S.Ct. 236, 50 L.Ed.2d 166 (1976). But the jury’s fact-finding power is not without limit. If, after drawing all reasonable inferences in favor of Northeastern, we are unable to conclude that the jury had a rational basis for its verdict, it is our duty to reverse the judgment entered on the jury verdict below. Michelman, supra, 534 F.2d at 1042. See Berkey Photo, Inc. v. Eastman Kodak Co., supra, 603 F.2d at 289. As previously discussed, four claims went to the jury: SNET was accused of (1) monopolizing and (2) attempting to monopolize the business terminal equipment market, and, in conjunction with AT&T and Western Electric, was alleged to have (3) conspired to monopolize and (4) conspired to restrain trade. Although at trial appellants contested every element of these four claims, they have now focused their attack on Northeastern’s assertions of anticompeti-tive conduct. Accordingly, we will assume that SNET possessed monopoly power in the contested market. We proceed to explain how, on the facts of this case, a conclusion that appellants’ activities were not anticompetitive would absolve them of liability under all four claims. 1. Monopolization Although monopoly power is the essence of a § 2 violation, [t]he mere possession of monopoly power does not ipso facto condemn a market participant. But, to avoid the proscriptions of § 2, the firm must refrain at all *85 times from conduct directed at smothering competition. This doctrine has two branches. Unlawfully acquired power remains anathema even when kept dormant. And it is no less true that a firm with a legitimately achieved monopoly may not wield the resulting power to tighten its hold on the market. Berkey Photo, Inc., supra, 603 F.2d at 275. See United States v. Grinnell Corp., 384 U.S. 563, 570-71, 86 S.Ct. 1698, 1703-04, 16 L.Ed.2d 778 (1966); SCM Corp. v. Xerox Corp., 645 F.2d 1195, 1205 (2d Cir. 1981). While Northeastern contends that the appellants’ monopoly power was “illegally obtained,” and depicts the Bell System as the epitome of the slothlful monopolist, it did not undertake to prove this contention to the jury. Instead, Northeastern vigorously asserts that, having become monopolists, by whatever means, appellants attempted to smother competition. Thus, the second branch of the Berkey holding is applicable: to succeed on its monopolization claim, Northeastern must show that SNET’s business practices were an exercise of its power over the market. 2. Attempt to Monopolize Anticompetitive or exclusionary conduct is unquestionably a necessary element of a § 2 attempt. See, e. g., Swift and Co. v. United States, 196 U.S. 375, 396-402, 25 S.Ct. 276, 279-282, 49 L.Ed. 518 (1905). See generally 3 P. Areeda and D. Turner, Antitrust Law 11820, 825-30 (1978). Moreover, the conduct requirement is arguably the single most important aspect of this offense. Proof of unlawful conduct may be used to infer specific intent to monopolize, see California Computer Products, Inc. v. International Business Machines Corp., 613 F.2d 727, 737 (9th Cir. 1979), and, when coupled with proof of monopoly power, it may satisfy the requirement that the attempt have a “dangerous probability of success,” see id. Because section 2’s prohibition of attempts to monopolize encompasses conduct by firms lacking monopoly power, L. Schwartz & J. Flynn, Antitrust and Regulatory Alternatives 12 & n.13 (5th ed. 1977), its potential reach is broader than the proscription against monopolization. But on the facts of this case, where SNET’s power to control prices and to exclude entry is no longer an open question, the two offenses are coterminous. To prove either violation, Northeastern must demonstrate that SNET engaged in anticompetitive behavior. 3. Conspiracy to Monopolize and to Restrain Trade The essence of these offenses is an agreement entered into with the specific intent of achieving monopoly or unreasonably restraining trade. United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 252, 60 S.Ct. 811, 857, 84 L.Ed. 1129 (1940); United States v. United States Gypsum Co., 600 F.2d 414, 417 (3d Cir.), cert. denied, 444 U.S. 884, 100 S.Ct. 175, 62 L.Ed.2d 114 (1979). The Sherman Act “does not make the doing of any act other than the act of conspiring a condition of liability.” Nash v. United States, 229 U.S. 373, 378, 33 S.Ct. 780, 782, 57 L.Ed. 1232 (1913). Thus, anticompetitive conduct is not indispensable in proving a § 1 or § 2 conspiracy. But it is frequently impossible for a plaintiff to obtain direct evidence of the alleged conspirators’ specific intent. In such situations, the finder of fact must be allowed to infer defendants’ intent from their anticompetitive practices. Such is the case here. Northeastern has not directed our attention to any unequivocal evidence of appellants’ specific intent to monopolize or unreasonably to restrain trade. Instead, it points to documents circulated among the corporations of the Bell System and to testimony of Bell employees. These show only that appellants wanted to win the competitive struggle. This desire, without more, is not unlawful. See Berkey Photo, Inc., supra. The crucial question is whether appellants specifically intended to vanquish their opposition by unfair or unreasonable means. Northeastern’s direct evidence is insufficient, by itself, to enable the jury rationally to conclude that appellants possessed the requisite intent. But because we are to draw all reasonable infer- *86 enees in appellee’s favor, see Continental Ore Co., supra, 370 U.S. at 696, 82 S.Ct. at 1409, we must uphold the verdict if we find that Northeastern presented circumstantial evidence of intent, i. e., that appellants engaged in anticompetitive conduct. Thus, on these claims as well, Northeastern can prevail only if appellants’ business practices are determined to be exclusionary. We come, at last, to grips with that question. B. The Evidence' I. Pricing We consider first the instances oí allegedly anticompetitive behavior that the district court properly described as the “heart of Northeastern’s case” — the pricing claims. Three such claims went to the jury: that appellants’ PBX and key telephone rates were predatory and that two-tier pricing was anticompetitive. 10 a. SNETs PBX rates Although the term “predatory pricing” lacks a precise economic meaning, see Bau-mol, Quasi-Permanence of Price Reductions: A Policy for Prevention of Predatory Pricing, 89 Yale L.J. 1, 26 & n.50 (1979), courts and commentators have generally defined predation as “the deliberate sacrifice of present revenues for the purpose of driving rivals out of the market and then recouping the losses through higher profits earned in the absence of competition.” 3 P. Agreeda & D. Turner, supra, H 711b, at 151. 11 See, e. g. International Air Industries, Inc. v. American Excelsior Co., 517 F.2d 714 (5th Cir. 1975), cert. denied, 424 U.S. 943, 96 S.Ct. 1411, 47 L.Ed.2d 349 (1976). Detailed economic analysis of this behavior is of comparatively recent vintage, gaining wide recognition only in 1975, with the publication of Areeda & Turner’s incisive article. See Areeda & Turner, Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harv.L. Rev. 697 (1975). This approach involves a comparison of a monopolist’s prices and expenditures, and necessarily entails an understanding of the various economic costs that confront a firm. These expenses fall into two rough categories — variable costs, those which fluctuate with a firm’s output, and fixed costs, those which are independent of output. Variable costs typically include such items as materials, fuel, labor, maintenance, licensing fees, and depreciation occasioned by use. The sum of all variable costs divided by output yields average variable cost. Fixed costs generally include management expenses, interest on bonded debt, the rate of return necessary to attract and maintain equity investment, irreducible overhead, and depreciation occasioned by obsolescence. 12 The sum of the firm’s fixed and variable costs divided by output equals average cost. By definition, average cost exceeds average variable cost at all levels of output. See 3 P. Areeda & D. Turner, supra, at UK 712, 715c. *87 Marginal cost, unlike the categories just defined, cannot be determined using data generated by conventional accounting methods; it is an economist’s construction. It is traditionally defined as “the increment to total cost that results from producing an additional increment of output.” Id. H 712, at 155. In most industries, marginal cost is low at low levels of output. It may decline slightly as output increases, but soon reaches a minimum, and then increases continuously with further increases in production. Thus, at low output levels, it is less than either average variable cost or average cost. At high levels, it is greater than either. 13 The legal question thus arises: what measure of cost should be used in determining whether a monopolist’s prices are unre-munerative, and hence predatory? This question must be approached with an understanding of the goals of antitrust law and an appreciation of the limits imposed by the judicial process. When Congress passed the Sherman Act in 1890, it “engraved in law a firm national policy that the norm for commercial activity must be robust competition.” Berkey Photo, Inc., supra, 603 F.2d at 272. Thus, should a conflict arise in a particular case between the desire to preserve the competitive process and the wish to rescue a competitor, courts must favor competition. Indeed, so pervasive is this principle that in spite of the law’s abhorrence of monopoly, even monopolists must not, without more, be flatly forbidden from competing. See id. at 273-75. Adopting marginal cost as the proper test of predatory pricing is consistent with the pro-competitive thrust of the Sherman Act. When the price of a dominant firm’s product equals the product’s marginal costs, “only less efficient firms will suffer larger losses per unit of output; more efficient firms will be losing less or even operating profitably.” 14 Areeda & Turner, supra, 88 Harv.L.Rev. at 711. Marginal cost pricing thus fosters competition on the basis of relative efficiency. Establishing a pricing floor above marginal cost would encourage underutilization of productive resources and would provide a price “umbrella” under which less efficient firms could hide from the stresses and storms of competition. Moreover, marginal cost pricing maximizes short-run consumer welfare, 15 since when *88 price equals marginal cost, consumers are willing to pay the expense incurred in producing the last unit of output. At prices above marginal cost, per contra, output is restricted, and consumers are deprived of products the value of which exceed their costs of production. Finally, in choosing among the various tests that commentators have proposed to detect predatory pricing, we must be mindful both of the limits of the judicial process and the realities of the marketplace. Courts occasionally err in applying even the clearest legal rules. The more complicated the standard, the greater the chance for misapplication. Especially when the costs of a misjudgment are high and the prevalence of the conduct the law seeks to deter is low, simpler rules are preferable. See generally Joskow & Klevorick, A Framework for Analyzing Predatory Pricing Policy, 89 Yale L.J. 213 (1979). So it is here. Predatory pricing is difficult to distinguish from vigorous price competition. Inadvertently condemning such competition as an instance of predation will undoubtedly chill the very behavior the antitrust laws seek to promote. Whether this risk is worth running depends in part of the prevalence of truly predatory conduct. There is considerable evidence, derived from historical sources and from economic teaching, that predation is rare. See 3 P. Areeda & D. Turner, supra, 1711a, at 150-52; McGee, Predatory Pricing Revisited, 23 J.L. & Econ. 289, 290-300 (1980). Indeed, nowhere in the recent outpouring of literature on the subject do commentators suggest that such pricing is either common or likely to increase. This does not mean, of course, that this behavior should no longer be deemed anticompetitive. But the rarity of the phenomenon informs our decision as to the appropriate legal definition. We agree with Areeda and Turner that in the general ease at least, the relationship between a firm’s prices and its marginal costs provides the best single determinant of predatory pricing. Thus, prices below reasonably anticipated marginal cost will be presumed predatory, 16 while prices above reasonably anticipated marginal cost will be presumed non-predatory. See 3 P. Areeda & D. Turner, supra, 1711d, at 153-54. And because marginal cost cannot be determined from conventional accounting methods, we will use average variable cost as its surrogate. Accord, Chillicothe Sand & Gravel Co. v. Martin Marietta Corp., 615 F.2d 427, 432 (7th Cir. 1980); Janich Brothers, Inc. v. American Distilling Co., 570 F.2d 848, 858 (9th Cir. 1977), cert. denied, 439 U.S. 829, 99 S.Ct. 103, 58 L.Ed.2d 122 (1978); Pacific Engineering & Production Co. v. Kerr-McGee Corp., 551 F.2d 790, 797 (10th Cir.), cert. denied, 434 U.S. 879, 98 S.Ct. 234, 54 L.Ed.2d 160 (1977); Hanson v. Shell Oil Co., 541 F.2d 1352, 1358 (9th Cir. 1976), cert. denied, 429 U.S. 1074, 97 S.Ct. 813, 50 L.Ed.2d 792 (1977); International Air Industries, Inc. v. American Excelsior Co., supra, 517 F.2d at 724. Were this a run-of-the-mill case we would proceed to apply the average variable cost rule to SNET’s PBX prices. But as Northeastern rightly contends, this case is far from typical. Indeed, two significant factors separate this from the classic situation: SNET offers more than one product and is regulated by the DPUC. That SNET is a multi-product firm seems to enhance the likelihood that it would prof *89 it from a policy of predatory pricing. One might assume that while SNET leased its PBXs at unremunerative rates, it would subsidize its losses with profits earned in other areas of its business. But although subsidization may stave off bankruptcy, it does not appreciably reduce the short-run costs of predation. In terms of lost profits, or in economic jargon “opportunity costs,” unremunerative prices will be as expensive to a diversified monopolist as to a single-product firm. Furthermore, the subsidization theory ignores an important prerequisite of a successful predatory campaign. For unremunerative pricing to make economic sense, the predator must be assured that he will be able to recoup his short term losses in the future. But because a dollar now is worth more than a dollar later (because of both inflation and the time value of money 17 ), he must be reasonably certain that once his prey has fallen, he will be able to reap supranormal returns. Thus it is not enough that the predator survive his own predatory prices; he must be able to earn monopoly profits in the not-too-distant future. Such profits, of course, will invite new entry. See, e. g., L. Sullivan, supra, at 117; 3 P. Areeda & D. Turner, supra, 1618a, at 41-42. Accordingly, in industries in which entry is easy, predatory pricing will not pay. As Northeastern’s own experience indicates, the barriers to entry into the business telephone equipment market were relatively low. 18 This does not mean, of course, that Northeastern could not have been a victim of predation. It merely shows that this market was not conducive to a policy of unremunerative pricing. Therefore, we need not adopt a more stringent test of predation merely because SNET is a diversified firm. 19 Cf. Joskow & Klevorick, supra, at 227-31. Northeastern pins its last hopes for an average or fully-distributed 20 cost test on the second distinguishing characteristic of this case — that SNET is subject to the authority of a state regulatory agency. Northeastern contends that even if the market can be relied upon to apportion the joint expenses of an unregulated firm, 21 a regulated utility can allocate all of its joint costs to the monopoly aspects of its business, thereby giving it a permanent advantage over its unregulated competitors. 22 *90 Judge Eginton accepted this argument, reasoning that: [t]o permit SNET to allocate all or a disproportionate percentage of its unattributable cost to the noncompetitive sector of its business ... is to give defendants an improper pricing advantage over Northeastern and other single market competitors who must allocate all costs directly to their terminal equipment rates. A fully distributed cost standard, on the other hand, would . . . substantially curtai[l] potential subsidization. This analysis is seriously flawed. First, it proves too much. The advantage Judge Eginton identifies may be enjoyed by all diversified firms, whether regulated or unregulated. But as we have previously discussed, allowing a diversified firm to maintain an artificial pricing floor above marginal cost is impractical and provides a haven for inefficient competitors. Second, this approach emphasizes the interests of single-market rivals over those of consumers and the competitive process. Marginal cost pricing maximizes short run consumer welfare. See 3 P. Areeda & D. Turner, supra, 1715a, at 168. Fully distributed cost pricing, in contrast, requires some consumers to pay a higher price for the desired product, and forces others to do without that product entirely, although they are willing to pay the costs of its production. Moreover, a fully distributed cost rule would have perverse consequences when the dominant firm’s rival was itself diversified. In that case, presumably, both firms would have to set their prices above their fully distributed costs. So high a price floor might prevent the diversified rival from ever entering the market, thereby curtailing competition rather than promoting it. Third, Northeastern’s argument in favor of the fully distributed cost test is based on a misunderstanding of the economic notion of subsidization. Northeastern seems to believe that whenever a product’s price fails to cover fully distributed costs, the enterprise must subsidize that product’s revenues with revenues earned elsewhere. But when the price of an item exceeds the costs directly attributable to its production, that is, when price exceeds marginal or average variable cost, no subsidy is necessary. On the contrary, any surplus can be used to defray the firm’s non-allocable expenses. Finally, Northeastern’s fear that SNET will be able to allocate all of its overhead to its monopoly services rests on the premise that the DPUC is either asleep or incompetent. A key step in the long and complex business of rate regulation is the “cost allocation study,” in which the regulatory agency distributes fixed costs against the revenues from the regulated and unregulated aspects of the utility’s operations. If Northeastern believes that the DPUC is unable to perform these studies, its recourse is to intervene before that body and, if unsuccessful, to appeal to the state courts. If comity and federalism mean anything in this context, they require that we not create an exception to the general rule of marginal cost pricing on the basis of plaintiff’s bald assertion that the DPUC cannot perform the duties delegated to it by the state. To summarize, we are unpersuaded by Northeastern’s contentions that fully distributed cost is the relevant standard against which to measure SNET’s PBX prices. Such a test would promote economic inefficiency, would be difficult to apply, would elevate the interests of single-market competitors over those of consumers, and would require us to assume that Connecticut is incapable of regulating telecommunications carriers. Having set forth our extensive exegesis to justify our selection of the marginal cost/average variable cost rule, we turn at last to an evaluation of Northeastern’s evidence of predatory pricing. Our treatment will be brief, since the record contains no evidence that SNET priced its PBXs below marginal cost. Northeastern’s expert witness, John Wilson, offered voluminous testimony concerning the various pric *91 ing strategies of SNET, AT&T, and Western Electric. He criticized those procedures at length and presented his opinion as to the proper method of pricing the Dimension 100 and the Dimension 400 PBXs. But all of his testimony was predicated on his assumption that SNET’s prices were required to exceed fully distributed costs. 23 Northeastern presented no evidence that SNET’s PBX prices were below marginal or average variable cost. Accordingly, we must reverse this portion of the judgment. We hasten to add, however, that pricing methods similar to the one SNET employed may, in other circumstances, run afoul of the antitrust laws. Had Northeastern proved that SNET omitted certain direct costs from its pricing studies, and that with the inclusion of those expenditures, marginal cost would have exceeded price, it would have established a prima facie case of predatory pricing. 24 b. The Two-Tier Payment Option The traditional monthly fees SNET’s customers paid for telephone service reflected both the capital costs of the equipment SNET provided and the operating expenses it incurred. Under this “permanent pay” plan, a customer continued making capital contributions for as long as he used equipment supplied by the utility. Thus, customers of long standing might, over time, pay an amount that exceeded SNET’s equipment costs. The new entrants into the interconnect market did not follow this customary method of revenue collection. Instead, they offered a variety of payment options, from outright purchases to leases of various kinds. To counter the erosion these resourceful offerings caused in its market position, SNET began promoting its own customized payment plan. This option, the so-called two-tier plan, was introduced as an alternative to SNET’s conventional billing method. A customer selecting the two-tier plan would agree to pay for capital equipment costs (Tier A) in equal monthly installments over a fixed period of time. The contractual period could range from one to ten years, depending upon the customer’s preferences and the type of equipment involved, but would cease at the end of the contractual period. A plan participant would pay SNET’s operating costs (Tier B) for as long as he used the equipment. Unlike the Tier A charge, the Tier B component could be increased with the approval of the DPUC. But, as stated, once the customer had completed the agreed-upon schedule of Tier A payments, he would pay only the Tier B assessment. If he preferred to end this arrangement be *92 fore the due date, he would be required to pay a termination charge equal to the discounted present value of his remaining Tier A payments. He would also receive a “termination credit” if SNET could use the equipment in providing service to another customer. Depending on the type of equipment and the original Tier A payment schedule, this credit might be greater or lesser than the termination charge. SNET did not make this option available to all of its business customers simultaneously. On the contrary, it introduced two-tier pricing in different segments of its business as they came under competitive attack. The first items to be offered on the two-tier plan were the larger models of SNET’s ComKey key telephone system. Two-tier pricing was later made available to SNET’s PBX customers, and finally, to users of a smaller ComKey system, the ComKey 416. Northeastern contended below, and Judge Eginton agreed, that SNET’s implementation of two-tier pricing was anticom-petitive. Northeastern’s argument was based on the testimony of John Wilson, its expert economist, who stated that this pricing scheme was an “economic lock-in.” In Wilson’s view, once a SNET customer began making Tier A payments, he would not switch to a competitor’s product, since to do so he would not only incur a termination charge but would also duplicate the capital charges he had already paid to SNET. The consequence of this payment scheme, Wilson concluded, was to remove a SNET customer from the marketplace for the full useful life of the equipment covered by the two-tier plan. Appellants now seek to refute the claim that two-tier pricing stifled competition, resting their argument on the following language in Berkey: [M]any anticompetitive actions are possible or effective only if taken by a firm that dominates its smaller rivals.. .. Such conduct is illegal when taken by a monopolist because it tends to destroy competition, although in the hands of a smaller market participant it might be considered harmless, or even “honestly industrial.” Berkey Photo, Inc. v. Eastman Kodak Co., supra, 603 F.2d at 274-75 (citations omitted; emphasis supplied). Appellants point out that many of SNET’s competitors offered payment plans that were virtually identical to two-tier pricing. For its part, Northeastern vigorously disputes appellants’ interpretation of the Berkey passage, insisting that it supports the view that this conduct was unlawful. To resolve this issue, therefore, we must explicate the Ber-key quotation. Before proceeding, some economic background is in order. As a preliminary matter, we question the conclusion of Northeastern’s expert witness that two-tier pricing removes a SNET customer from the market for a significant period of time. This view ignores an important feature of SNET’s two-tier plan — the termination credit. Because of this credit, a SNET customer wishing to purchase a competitor’s equipment will not lose all of his previous investment in SNET’s product. He can, in effect, “resell” his present equipment to SNET 25 and apply that money to the price of the competitor’s service offering. That the termination credit may only reduce the termination charge, instead of producing a net refund, does not undermine its financial significance. It is reasonable to suppose that one who has just made his first Tier A payment is unlikely to buy a Northeastern PBX. But the incentives influencing his decision are no different than those operating on one who has just bought any other consumer item, a new car, for example. Northeastern’s expert would have us believe that a consumer will not purchase a newer model until after the old one has worn out. Common experience belies this view. *93 The foregoing analysis demonstrates that the incentives operating on a SNET customer who elects the two-tier plan are similar to those applicable to a new owner of a Northeastern PBX. Just as a SNET customer is unlikely to purchase Northeastern’s product, the latter is unlikely to switch to SNET’s terminal equipment. Thus, SNET’s use of the two-tier plan is no more anticompetitive than Northeastern’s practice of selling PBXs. Neither one is an action that is “possible or effective only if taken by a firm that dominates its smaller rivals.” Berkey, supra, 603 F.2d at 275. On the contrary, both are “ordinary marketing methods available to all in the market.” Telex Corp. v. International Business Machines Corp., 510 F.2d 894, 926 (10th Cir.), cert. dismissed, 423 U.S. 802, 96 S.Ct. 8, 46 L.Ed.2d 244 (1975). Thus, we conclude that appellants’ two-tier payment option was not anticompetitive, and that this portion of the jury’s verdict cannot stand. 2. Advertising To avoid prolonging a necessarily extended opinion, we will consider briefly the appellants’ remaining challenges to the sufficiency of Northeastern’s evidence. At trial, appellee attempted to prove that SNET’s advertising was excessive and disparaged Northeastern’s products. Since the disparagement claim did not go to the jury, we need not consider it further. As for the remaining allegation, Northeastern demonstrated that in 1969, shortly after the FCC’s Carterfone decision, SNET spent only $100 for advertising, but that by 1978, this expenditure had risen to $450,000, an amount representing roughly V2 of 1% of SNET’s related revenue for this time period. Berkey teaches that “[a] monopolist is not forbidden to publicize its product unless the extent of this activity is so unwarranted by competitive exigencies as to constitute an entry barrier.” 603 F.2d at 287. Northeastern presented no evidence from which the jury could reasonably infer that SNET’s advertising was either unwarranted or that it significantly raised the barriers to entry. Accordingly, this portion of the judgment must also be reversed. 3. Product Introduction and Marketing We will consider these claims together since they are based on a common fallacy. Northeastern suggests that a monopolist’s efforts in meeting competition are to be condemned unless they are intended to promote consumer welfare. While it is a fundamental tenet of antitrust law that customers will benefit from the salubrious effects of competition, a monopolist’s right to compete is not limited to actions undertaken with an altruistic purpose. Even monopolists must be allowed to do as well as they can with their business. Buffalo Courier-Express, Inc. v. Buffalo Evening News, Inc., 601 F.2d 48, 55 (2d Cir. 1979). In light of these by-now-familiar principles, it is apparent that we must exercise considerable caution in condemning SNET merely for introducing new products, see Berkey, supra, 603 F.2d at 286 & n.30, or for reorganizing its marketing operations, see Sargent-Welch Scientific Co. v. Ventron Corp., 567 F.2d 701 (7th Cir. 1977), cert. denied, 439 U.S. 822, 99 S.Ct. 87, 58 L.Ed.2d 113 (1978). While situations may arise in which such actions would violate the Sherman Act, 26 our thorough examination of appellee’s evidence indicates that such a case is not presented here. Consequently, these portions of the judgment must be reversed. 4. SNET’s Use of Its Utility Function Any business choosing to purchase Northeastern’s terminal equipment was still dependent on SNET for local and long distance telephone service. Northeastern asserts that SNET unreasonably impeded competition in the terminal equipment market by exercising its monopoly power over the provision of these services. In particular, appellee maintains that SNET attempt *94 ed to discourage businesses from acquiring Northeastern’s equipment by intentionally providing poor service to Northeastern’s customers. But once again, appellee’s proof cannot pass muster even under the generous 'Continental Ore standard. Northeastern did not introduce any evidence whatsoever that SNET provided appreciably poorer service to Northeastern’s patrons than it did to businesses using SNET equipment. Similarly, there was not a scintilla of evidence that SNET’s failure to provide better service to plaintiff’s customers was intentional. Northeastern merely demonstrated that SNET often underestimated the time necessary to cure various electrical malfunctions, including “changeover” problems — /. e., those arising when Northeastern’s equipment was connected into SNET’s telecommunications system. This evidence was insufficient to sustain a finding that SNET deliberately acted to stifle competition. Cf. George R. Whitten, Jr., Inc. v. Paddock Pool Builders, Inc., 508 F.2d 547 (1st Cir. 1974), cert. denied, 421 U.S. 1004, 95 S.Ct. 2407, 44 L.Ed.2d 673 (1975). This portion of the jury’s verdict also cannot stand. 5. Protective Couplers We arrive at last at the one area in which Northeastern presented some evidence of appellants’ anticompetitive conduct. As noted previously, Northeastern does not challenge the lawfulness of the tariffs AT&T filed immediately following the Carterfone decision. It does contend, however, that AT&T intentionally designed the protective couplers to impede competition in the terminal equipment market. Specifically, appellee objects that the couplers’ leads were incompatible with its PBXs and that the couplers required an external power source. 27 In support of its design defect claims, Northeastern relied heavily on the conclusions reached in a National Academy of Sciences study which the FCC had commissioned. Although the NAS panel found AT&T’s coupler design to be an “adequate” solution to the problems of interconnection, it also recommended, as we noted above, that the FCC consider a registration system whereby equipment meeting certain minimum standards would be exempt from the coupler requirement. Of particular relevance to Northeastern’s allegations, the panel concluded that the couplers’ dependence on commercial power was “a significant and probably unnecessary disadvantage.” This judgment finds some support in the testimony of James L. Simon, a Bell electrical engineer, who admitted on cross-examination that the coupler could have been designed to operate using the power available in SNET’s telephone lines. Thus, Northeastern’s claim that appellants intentionally designed the protective coupler in an unreasonable manner is not totally unsubstantiated. Nevertheless, we cannot uphold the jury’s findings of liability. As mentioned previously, the jury’s decisions on the four antitrust claims were contained in its answers to fourteen interrogatories. While the district court’s third query asked the jury whether each of six activities was anticompetitive, succeeding questions, which were crucial to the jury’s verdict, did not differentiate among appellants’ allegedly exclusionary practices. For example, Question Four asked the jury if SNET’s “anticompetitive or predatory acts ... were the proximate cause of any injury or damage to [Northeastern’s] business or property.” Similarly, Questions Nine and Twelve, which related to the conspiracy claims, referred generally to appellants’ allegedly anticompetitive behavior. The jury’s affirmative responses to such questions do not tell us which practices formed the basis of its verdict. Had we found that all of appellants’ challenged conduct was anticompetitive, the drafting of the interrogatories would not have posed a problem. But Northeastern failed to prove that five of those six activities were exlusionary. Because we cannot be certain that the jury *95 based its verdict exclusively on the one remaining practice, without any spill-over from the practices we have found not to have been anticompetitive, 28 its findings of liability must be set aside, and a new trial ordered. Moreover, even if we could determine that the jury would have found appellants liable based solely on the design of the protective coupler, a new trial on both liability and damages would still be necessary. In Gasoline Products Co. v. Champlin Refining Co., 283 U.S. 494, 51 S.Ct. 513, 75 L.Ed. 1188 (1931), the Supreme Court addressed the propriety of ordering a new trial limited to a redetermination of damages. The Court stated that a partial new trial “may not properly be resorted to unless it clearly appears that the issue to be retried is so distinct and separable from the others that a trial of it alone may be had without injustice.” Id. at 500, 51 S.Ct. at 515, 75 L.Ed. 1188. See also Caskey v. Village of Wayland, 375 F.2d 1004, 1009-10 (2d Cir. 1967); 6A Moore’s Federal Practice 159.06 (2d ed. 1976). In the instant case, the proper calculation of Northeastern’s lost profits will involve a comparison of the profits Northeastern actually earned with the profits it would have earned had appellants supplied its customers with reasonably designed protective couplers. Thus, the reasonableness of the couplers’ design will be a crucial issue at both the liability and damages phases of any further proceedings. Here, as in Gasoline Products Co., “the question of damages ... is so interwoven with that of liability that the former cannot be submitted to the jury independently of the latter without confusion and uncertainty, which would amount to a denial of a fair trial.” Id at 500, 51 S.Ct. at 515, 75 L.Ed. 1188. To prevent an injustice to the parties, we remand this case for a new trial at which Northeastern may attempt to prove that appellants’ design of the protective coupler violated the Sherman Act. 29 Should this endeavor prove successful, the jury may award Northeastern appropriate compensation. 30 *96 IV. CONCLUSION Because the issues raised on this appeal are numerous, we briefly summarize our disposition: (1) Of the six types of conduct Northeastern alleges to have been anticom-petitive, we find that it failed to prove five — pricing, advertising, product introduction, marketing, and SNET’s use of its utility function. (2) Although Northeastern presented some evidence that appellants intentionally designed their protective couplers to impede competition in the business terminal equipment market, we must set aside the jury’s verdict of liability. Because the interrogatories submitted to the jury did not permit it to differentiate with sufficient precision among the six types of allegedly anticompetitive behavior, we cannot be certain that the jury based its verdict solely on the evidence concerning the couplers’ design. Accordingly, a new trial is ordered to enable Northeastern to prove — without resort to evidence of conduct that we have found not to have been anticom-petitive — that appellants’ design of the protective coupler violated the Sherman Act. (3) Since we have reversed the judgment for Northeastern in major respects, the award of attorneys’ fees is vacated and remanded to the district court for reconsideration in light of our holdings. Footnotes [1] The Bell System consists of AT&T, Western Electric, Bell Laboratories, and the twenty-three affiliated operating companies that provide local telephone service. [2] SNET is one of two local affiliates in which AT&T possesses a minority interest. [3] SNET, which is specially chartered by the Connecticut General Assembly and is under the regulatory supervision of the Connecticut Division of Public Utilities Control (DPÜC), does not sell or lease equipment to its customers; it provides services pursuant to tariffs. For convenience, however, we will refer to SNET’s service offerings as “products.” [4] These numerical designations refer to the number of telephone lines each type of PBX could accommodate. [5] As a consequence of this design feature, a Northeastern customer, unlike one using SNET equipment, would lose outside telephone service in the event of a power failure. [6] The parties have stipulated that this is the relevant product and geographic market for the purposes of this litigation. [7] Appellants also maintain that the remaining areas of conduct challenged by Northeastern— pricing, advertising, marketing, the introduction of new products, and the use of SNET’s utility function — are insulated from antitrust scrutiny by the “state action” doctrine first enunciated in Parker v. Brown, 317 U.S. 341, 63 S.Ct. 307, 87 L.Ed. 315 (1943). While there is considerable force to the contentions of Northeastern and of the Government as amicus curiae that such immunity is unavailable, see Can tor v. Detroit Edison Co., 428 U.S. 579, 96 S.Ct. 3110, 49 L.Ed.2d 1141 (1976); Goldfarb v. Virginia State Bar, 421 U.S. 773, 95 S.Ct. 2004, 44 L.Ed.2d 572 (1975); Litton Systems, Inc. v. American Telephone & Telegraph Co., 487 F.Supp. 942, 958-59 (S.D.N.Y.1980), we need not reach this issue in view of our holding that on the facts of this case, none, of the charged activities gives rise to antitrust liability. See Part III. B. infra. [8] At that time, the Federal Communications Act empowered the FCC to suspend a proposed tariff for only ninety days. See 47 U.S.C. § 204, as amended, Pub.L.No.94-376, § 2, 90 Stat. 1080 (1976). The maximum period of suspension has since been increased to five months. See id. [9] The potential for conflict between the regulatory and antitrust provisions is further reduced because Northeastern seeks damages only for appellants’ past activities. Had it requested an injunction prohibiting appellants from engaging in conduct expressly approved by the FCC, a different question might be presented. [10] The jury was not asked to consider these issues separately. Judge Eginton, in ruling on appellants’ motion for judgment notwithstanding, the verdict, concluded there was insufficient evidence that the key telephone rates were below “cost.” The court nevertheless refused to set aside the verdict, professing certainty that the jury did not base its finding of liability on that evidence. In view of our disposition of the other two pricing claims, we need not decide whether this conclusion was erroneous. Moreover, because Northeastern did not cross-appeal from the judgment below, the issue of the lawfulness of those rates is not before us. [11] Predation may also refer to a monopolist’s practice of “selling at a loss” to discipline smaller rivals for undercutting the monopoly price. See 3 P. Areeda & D. Turner, Antitrust Law U 711a, at 151 n.5 (1978). As in the first case, the predator expects to recoup his losses once the disciplinary period is over. Northeastern has not accused appellants of this type of predatory conduct, acknowledging, presumably, that because they are subject to regulation, they lack the pricing flexibility necessary to respond quickly to the price reductions of rebellious firms. [12] Whether a particular expense, e. g., the cost of a new factory, should be classified as variable or fixed depends in part on the time period under consideration. Over the short run, such an expense would be independent of output and hence fixed. In the long run, it would be classified as variable. Indeed, all costs are variable in the long run. See 3 P. Areeda & D. Turner, supra, (] 712, at 155-56. [13] The relationships among marginal cost, average cost, and average variable cost are depicted graphically in Areeda & Turner, Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harv.L.Rev. 697, 701 n.14 (1975). [14] It may seem incongruous that a more efficient firm may lose money at the monopolist’s lawful price, see In re International Business Machines Corp. Peripheral Electronic Data Processing Devices Antitrust Litigation, 481 F.Supp. 965, 991-95 (N.D.Cal.1979) (dictum); Comment, Draining the Alcoa “Wishing Well": The Section 2 Conduct Requirement after Kodak and Calcomp, 48 Ford.L.Rev. 291, 317-20 (1979). But these losses are unlikely to have any serious anticompetitive effects. If the monopolists’ marginal-cost price is below its average cost, not only will it be incurring greater losses per unit of output than the more efficient firm, but it will suffer those losses over a much greater range of output, since it will be making the vast majority of sales in the market. That the monopolist will suffer more than its rival will tend to negate whatever competitive benefits the monopolist may derive from its presumably superior wealth. Moreover, since the monopolist will have to raise his price eventually, the rival has a strong incentive to stay in the market, because its profits will exceed those of the monopolist at the new, higher price. See generally McGee, Predatory Pricing Revisited, 23 J.L. & Econ. 289, 295-97 (1980). Finally, the-average cost rule that some commentators would apply in this situation has serious flaws of its own. See note 19 infra. [15] Areeda & Turner’s concern over maximizing short term as opposed to long term welfare has provoked considerable criticism. See, e. g., Scherer, Predatory Pricing and the Sherman Act: A Comment, 89 Harv.L.Rev. 869 (1976); Scherer, Some Last Words on Predatory Pricing, id. at 901; Williamson, Predatory Pricing: A Strategic and Welfare Analysis, 87 Yale L.J. 284 (1977). Maximizing consumers’ welfare over the long run is the theoretically preferable solution, as Areeda and Turner concede. See Areeda & Turner, Scherer on Predatory Pricing: A Reply, 89 Harv.L.Rev. 891, 896-97 (1976). But we despair that economists will ever formulate workable criteria for reaching this goal. Certainly Scherer’s approach cannot be applied in a judicial setting. See Janich Brothers, Inc. v. American Distilling Co., 570 F.2d 848, 857 n.9 (9th Cir. 1977), cert. denied, 439 U.S. 829, 99 S.Ct. 103, 58 L.Ed.2d 122 (1978); McGee, supra, at 306-07. *88 Furthermore, this case does not present us with a stark choice between long run and short run welfare maximization, since SNET did not set its prices solely with respect to short term marginal costs. Instead it took such “long term” items as the return on investment into consideration in formulating its PBX tariffs. See note 19 infra. [16] Areeda and Turner would create an exception to the prohibition of below marginal cost pricing when a firm’s prices exceed its average costs, a situation that can only arise when the firm is operating at output levels above capacity. This exception has been strongly criticized. See Scherer, Predatory Pricing and the Sherman Act: A Comment, 89 Harv.L.Rev. 869, 871-75 (1976). Fortunately, we need not resolve this highly technical issue, since Northeastern does not allege that SNET’s marginal costs were above its average costs; it alleges the reverse. [17] See V. Brudney & M. Chirelstein, Corporate Finance (1972). [18] As mentioned previously, Northeastern was formed by two Connecticut businessmen in 1972 on a total capital investment of $1,000. [19] Moreover, the more demanding average cost test of predatory pricing gives rise to serious problems when applied to a diversified enterprise. This is because the accounting conventions used to allocate joint costs, i. e., those costs that are not directly attributable to a particular product, are all arbitrary to a degree. See 3 P. Areeda & D. Turner, supra, j| 719, at 183. Even Northeastern’s expert witness, who testified that appellants should have allocated a portion of their joint expenditures to the cost of their PBXs, did not explain how those costs should be apportioned. Since businessmen are entitled to notice that their pricing decisions may subject them to antitrust liability, they should be allowed to adopt any reasonable and consistent method for allocating joint expenditures. See In re IBM Peripheral EDP Devices Antitrust Litigation, supra, 481 F.Supp. at 998-99. SNET’s method, which it termed Long Range Incremental Analysis (LRIA), was to calculate all of the costs directly attributable to a particular product, including such items as operating expenses, capital costs, start-up charges, inflation factors, and “elastic costs” (i. e., the losses in revenue occasioned when a SNET customer switched from, using older SNET equipment to a new SNET product). SNET would then estimate the net revenues to be expected from pricing the new product at each of four or more price levels, and select the price that maximized net revenues. In essence, Long Range Incremental Analysis let the market allocate SNET’s joint costs. While this method is not beyond challenge, see pp. 89, 90, infra, we believe it was a reasonable procedure for apportioning common expenditures. [20] Judge Eginton defined fully distributed cost as the “marginal or incremental cost of supplying a particular good or service [plus] some portion of the unattributable costs (f. e., indirect costs.).” [21] See note 19 supra. [22] If an unregulated, diversified monopolist were to try this tactic, its high price on the product bearing all of the firm’s joint costs would come under heavy competitive attack. *90 To remain competitive, the monopolist would have to lower that product’s price, thereby reallocating common expenditures. [23] Wilson’s analysis of SNET’s first year “tracking report” does not cure this lack of evidence. During the first year that the Dimension PBXs were on the market, SNET compiled data on the costs and revenues associated with those products. That study, which covered the period from March 1, 1977, to February 28, 1978, showed that while Dimension’s revenues were greater than expected, so were its costs, resulting in a $2,000 shortfall where a $25,700 surplus had been forecast. But the existence of this loss does not demonstrate that SNET’s tariffs did not exceed its reasonably anticipated average variable costs. Two factors support this conclusion. The first involves the size of the loss itself. Even monopolists cannot be held to a standard of perfection. We have stated that SNET was entitled to price its PBXs at a level equal to its short run average variable costs. A loss of $2,000 on total revenues of more than $232,000 cannot subject SNET to antitrust liability under this standard. Moreover, both the tracking study and SNET’s Long Range Incremental Analyses, see note 19, su pra, include certain items of long term expense that are generally excluded from a calculation of short run average variable costs. See generally 3 P. Areeda & D. Turner, supra, j] 715c, at 173. For example, SNET included the rate of return on investment in estimating the costs of its Dimension PBX program. If this component were eliminated, the tracking study would have shown a profit instead of a deficit for the first year of the program’s operation. [24] Of course, SNET would still be entitled to prove that its below-cost prices had an acceptable business justification. See Hanson v. Shell Oil Co., 541 F.2d 1352, 1359 n. 6 (9th Cir. 1976), cert. denied, 429 U.S. 1074, 97 S.Ct. 813, 50 L.Ed.2d 792 (1977). One affirmative defense would be that its prices, while below marginal cost, were loss minimizing. Such prices are inconsistent with the intent to forego short term revenues — a necessary element of the offense of predatory pricing. See 3 P. Areeda & D. Turner, supra, ([711b, at 151. [25] Of course, if the customer knows SNET will not give him fair value for the equipment he has been using, he will be less inclined to trade it in. But in the absence of evidence that SNET regularly undervalued its customers’ used terminal equipment, we must assume that this practice did not occur. [26] For example, introduction of a new product may violate § 2 if a monopolist acts to compel customer choice by withdrawing a substitute product from the market. See Berkey, supra, 603 F.2d at 287 & n.39. [27] Although Northeastern also contended that its customers experienced problems concerning the supply and service of the protective cou-piers, those issues were not submitted to the jury, and we need not consider this evidence. [28] Northeastern argues strenuously that it “should be given the full benefit of [its] proof without tightly compartmentalizing the various factual components and wiping the slate clean after scrutiny of each.” Continental Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 699, 82 S.Ct. 1404, 1410, 8 L.Ed.2d 777 (1962). Although the logical presentation of our holding has required us to treat each of Northeastern’s claims of anticompetitive conduct separately, we have not ignored Continental Ore’s edict. We have found, however, that in numerous critical respects Northeastern’s proof was utterly lacking. Under such circumstances, treating Northeastern’s claims collectively cannot have any synergistic effect. See California Computer Products, Inc. v. International Business Machines Corp., 613 F.2d 727 (9th Cir. 1979). [29] In other circumstances, we might be reluctant to allow a jury to second-guess engineers’ decisions as to the proper construction of a sophisticated piece of equipment. But in this case we cannot look to the reaction of the competitive market to determine whether one design is superior to another, compare Berkey, supra, 603 F.2d at 287. As noted previously, ATT’s post-Carterfone tariffs gave it the exclusive right to provide protective couplers. Market forces cannot operate under such circumstances. Thus, we see no alternative to entrusting the matter of coupler design to the judgment of the jury. [30] To forestall a controversy that will doubtlessly recur should a second damage trial become necessary, we will consider appellants’ contention that Northeastern cannot recover both lost profits and damages to its going concern value. We recognize that these two types of recovery may be awarded in the same case, albeit not for the same time period. See Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 51 S.Ct. 248, 75 L.Ed. 544 (1931); Atlantic City Electrical Co. v. General Electric Co., 226 F.Supp. 59, 61 (S.D.N.Y.), application for an interlocutory appeal denied, 337 F.2d 844 (2d Cir. 1964) (per curiam). For example, recovery on both grounds will often be appropriate when an antitrust plaintiff can prove that he first lost business and was then driven from the market by the defendant’s anti-competitive practices. See, e. g., Heatransfer Corp. v. Volkswagenwerk, A.G., 553 F.2d 964, 987 n.20 (5th Cir. 1977), cert. denied, 434 U.S. 1087, 98 S.Ct. 1282, 55 L.Fd.2d 792 (1978). But Northeastern has not shown that its future profitability will be diminished by the design of appellants’ protective coupler. Indeed, to do so would be impossible, since the jury’s implicit finding that the couplers were unreasonably designed involved the ephemeral and now-invalidated post-Carterfone tariffs. Furthermore, *96 we are told that the couplers still in use have been modified to accommodate two-wire interconnection. Under these changed circumstances, Northeastern cannot recover on the theory that it suffered a loss to its value as a going concern. Cf. Volasco Products Co. v. Lloyd A. Fry Roofing Co., 308 F.2d 383, 393 (6th Cir. 1962), cert. denied, 372 U.S. 907, 83 S.Ct. 721, 9 L.Ed.2d 717 (1963).