Full opinion text
Opinion for the Court filed by Circuit Judge WALD. TABLE OF CONTENTS Page I. Statutory and Factual Background____ 300 A. The Statutory Scheme------------- 300 1. Specific Provisions Governing Railroad Mergers _________________ 300 2. Congressional Policy on Railroad Mergers _____________________ 301 B. ICC Railroad Merger Policy_________ 302 C. The Railroads Involved in this Case___ 303 D. Proceedings Below________________ 303 1. Conditions Requested by the Boston & Maine’s Competitors__________ 303 2. The ICC’s Decision_____________ 304 E. Issues Presented__________________ 304 II. The Financial Viability of the Boston & Maine............................. 305 III. Lamoille Valley’s Request for Protective Conditions........... 307 A. Standard of Review_______________ 307 B. The ICC’s Decision________________ 308 C. The Essential Services Test_________ 309 1. Current or Past Truck Service____ 310 2. Business Termination___________ 310 3. Conclusion____________________ 313 D. Anticompetitive Effect ____________ 313 IV. Canadian National’s Request for Protective Conditions______________________ 315 A. The ICC’s Decision_________________ 315 B. The Limited Relevance of Systemwide Revenues________________________ 316 C. Guilford’s Intent to Maintain Present Service-------- 317 D. Guilford’s Incentives to Downgrade Interchange Service_______________ 318 1. Guilford’s Incentives in Antitrust Theory ---------------------- 318 2. Guilford’s Actual Incentives______ 319 E. Canadian National’s Competitive Leverage _____________________________ 320 F. Potential Harm from Protective Conditions _____ 321 G. Conclusion_______________________ 322 V. Labor Protective Conditions.......... 323 VI. Procedural Issues.....--............ 324 A. Must Mr. Mellon Join the Merger Application? ________________________ 324 B. The Expedited Procedural Schedule .. 326 C. ICC Jurisdiction over Holding Company Securities________________________ 329 D. Failure to Apply for Control of the Vermont & Massachusetts Co._______ 330 E. Premature Control of the Boston & Maine..................-....... 330 VII. Conclusion ......................... 331 WALD, Circuit Judge: We review here a decision of the Interstate Commerce Commission (ICC or Commission) approving unconditionally the merger of the Maine Central Railroad with the Boston & Maine Railroad. Guilford Transportation Industries — Control—Boston & Maine Gorp., 366 I.C.C. 292 (1982) [hereinafter cited as Boston & Maine Merger]. Petitioners, competitors of the Boston & Maine, asked the ICC to protect them from competitive harm due to the merger by imposing various conditions on the merged entity (including sale of track, trackage rights, and preservation of swift traffic interchanges). The ICC declined to impose any protective conditions, finding that none of the petitioners had shown that the conditions it requested were needed to prevent the loss of “essential services.” Petitioners appeal to this court, claiming that the ICC’s “essential services” test for imposing protective conditions is too strict and does not comply with the statutory directive that the ICC consider the effect of the merger on “adequacy of transportation to the public.” 49 U.S.C. § 11,344(b)(1). Petitioners also argue that some of the ICC’s findings are not supported by substantial evidence and that the ICC committed a variety of procedural errors. We reject the procedural challenges as either without merit or not constituting prejudicial error. We conclude, however, that the ICC’s essential services test, as applied to petitioner Lamoille Valley Railroad, does not comport with the statute. We also find flaws in the agency’s reasoning with regard to protective conditions requested by petitioner Canadian National Railway. We therefore affirm in part, reverse in part, and remand to the ICC to determine whether protective conditions are needed to protect the public’s right to adequate transportation. I. Statutory and Factual Background A. The Statutory Scheme 1. Specific Provisions Governing Railroad Mergers Interstate Commerce Act, 49 U.S.C. § 11,343(a), requires the ICC to review all railroad mergers. The ICC must approve any merger that is “consistent with the public interest,” but can impose conditions on the merger when needed to advance the public interest: The Commission shall approve and authorize a [merger] when it finds the transaction is consistent with the public interest. The Commission may impose conditions governing the transaction. Id. § 11,344(c). In making the “public interest” determination for a proposed merger of two “class I” railroads, the ICC must consider five specific (but not exclusive) factors: (A) the effect of the proposed transaction on the adequacy of transportation to the public. (B) the effect on the public interest of including, or failing to include, other rail carriers in the area involved in the proposed transaction. (C) the total fixed charges that result from the proposed transaction. (D) the interest of carrier employees affected by the proposed transaction. (E) whether the proposed transaction would have an adverse effect on competition among rail carriers in the affected region. Id. § 11,344(b)(1). 2. Congressional Policy on Railroad Mergers These statutory provisions must be interpreted in light of the longstanding congressional policy favoring railroad mergers that increase efficiency and quality of service. Since 1920, the ICC has operated under “the congressional policy of encouraging consolidation of the Nation’s railroads.” Penn-Central Merger & N & W Inclusion Cases, 389 U.S. 486, 492, 88 S.Ct. 602, 605, 19 L.Ed.2d 723 (1968). The Railroad Revitalization and Regulatory Reform Act of 1976 (“4R Act”), Pub.L. No. 94-210, 90 Stat. 31 (current version in scattered sections of 45, 49 U.S.C.), reinforces that policy. The 4R Act did not alter the substantive “public interest” standard for ICC approval of railroad mergers, but did provide new, expedited procedures for ICC review of merger proposals as an alternative to the existing procedures. It also declared Congress’ purposes to include: the encouragement of efforts to restructure the [rail] system on a more economically justified basis, including ... an expedited procedure for determining whether merger and consolidation applications are in the public interest. Id. § 101(a)(2), 45 U.S.C. § 801(a)(2). The Senate committee report explained that the new procedures were needed because “[t]he cumbersome, slow process of ... processing merger proposals ... has drastically slowed change needed in the industry.” S.Rep. No. 499, 94th Cong., 1st Sess. 2-3 (1975), reprinted in 1976 U.S.Code Cong. & Ad.News 14, 16. The committee expected the new procedures to “encourage mergers, consolidations and joint use of facilities that tend to rationalize and improve the Nation’s rail system.” Id. at 20, 1976 U.S.Code Cong. & Ad.News at 34. Similarly, the House expected that the new procedures would permit railroads “to voluntarily rationalize the system in a short period of time.” H.R.Rep. No. 725, 94th Cong., 1st Sess. 63 (1975). The Staggers Rail Act of 1980, Pub.L. No. 96-448, 94 Stat. 1895 (amending 49 U.S.C. §§ 10,101-11,917), also shows Congress’ endorsement of mergers that enhance railroad efficiency. The Staggers Act did not change the Interstate Commerce Act provisions governing mergers of two or more class I railroads. However, it substantially deregulated the railroad industry in a variety of ways (notably rate-setting practices) and established “the policy of the United States Government ... to minimize the need for Federal regulatory control over the rail transportation system” and “to foster sound economic conditions in transportation.” 49 U.S.C. § 10,101a(2), (5). B. ICC Railroad Merger Policy In response to the congressional directives in the 4R Act and the Staggers Rail Act, the ICC has established policy guidelines which favor railroad mergers that improve efficiency and disfavor imposing conditions on a merger that might reduce potential efficiency gains. 49 C.F.R. § 1180.1 (1982); see Railroad Consolidation Procedures, 363 I.C.C. 200 (1981) (statement of basis and purpose). In general, the Commission “encourages ... rationalization of the nation’s rail facilities and reduction of its excess capacity.” 49 C.F.R. § 1180.1(a) (1982). In deciding whether a merger is in the public interest, the Commission “performs a balancing test,” weighing “the potential benefits to applicants and the public against the potential harm to the public.” Id. § 1180.1(c). In weighing potential harm to the public, the ICC considers two principal factors— the degree of reduction of competition from a merger and the harm, if any, to “essential services.” In determining whether competition will be significantly reduced, the Commission considers both “intramodal” competition among rail carriers and “inter-modal” competition with trucks, barges, pipelines, and ships. Id. § 1180.1(c)(2)(i). The ICC defines rail service as “essential” if “there is a sufficient public need for the service and adequate alternative transportation is not available.” Id. § 1180.-l(c)(2)(ii). The Commission does not take into account potential harm to individual competitors (except insofar as harm to competitors significantly reduces competition) because it is concerned only with rail service to the public, “not the survival of particular carriers.” Id. § 1180.1(c)(2)(ii). The ICC believes that protective conditions are generally not in the public interest because they “may lessen the benefits of a consolidation to both the carrier and the public.” Id. § 1180.1(d)(1). It recognizes that competition creates an incentive to provide efficient service and therefore will impose conditions when needed to “amelio-rat[e] potential anticompetitive effects of a consolidation.” But it “will not normally impose conditions” on a merger “to protect a carrier" unless the conditions are needed to preserve “essential services” and will neither impose “unreasonable” costs on the consolidated carrier nor “frustrate the ability of the consolidated carrier to obtain the anticipated public benefits.” Id. C. The Railroads Involved in this Case Guilford Transportation Industries (Guil-ford) is a holding company that is 100% owned by Timothy Mellon. Guilford owns the Maine Central Railroad, a profitable railroad with a near-monopoly over rail service to the Maine paper and forest products industries. The bulk of the traffic carried by the Maine Central originates in Maine and travels west (to Montreal, Buffalo, Chicago, etc.) or south (to Boston, New York, and the mid-Atlantic states); there are also significant shipments of grain from the Midwest to the Maine poultry industry. Rail traffic that originates on the Maine Central can move south over the Boston & Maine Railroad. Traffic from the Maine Central can move west over four routes: over the Canadian National Railway to Montreal and points west; over the Canadian Pacific Railroad to Montreal and points west; over the Lamoille Valley Railroad to the Canadian National to Montreal and points west; and over the Boston & Maine to Albany and from there via Conrail or the Delaware & Hudson Railroad to Buffalo and points west. The Canadian National and the Canadian Pacific are large and profitable transcontinental railroads. The Lamoille Valley operates a single 98-mile east-west line between Swanton, Vt. (its western gateway, connecting to the Canadian National) and St. Johnsbury, Vt. (its eastern gateway, connecting to the Maine Central). Only a small volume of traffic originates or terminates on Lamoille Valley’s line. Thus, it depends on “overhead” traffic (carried from the Maine Central over the Lamoille Valley to the Canadian National or vice-versa) for financial viability. Even with its current level of overhead traffic, the Lamoille Valley has been only marginally profitable, and then only because the State of Vermont spent some $16 million during the 1970’s in upgrading Lamoille Valley’s track and equipment. The Boston & Maine operates east-west and north-south rail lines in Vermont, New Hampshire, Massachusetts, Connecticut, and New York. It is bankrupt and has been in reorganization since 1970. The Boston & Maine has lost money for the last 25 years; its revenue insufficient to support its extensive network of rail lines. The Providence & Worcester is a small, profitable railroad serving Rhode Island and Connecticut. It connects primarily with the Boston & Maine and the Canadian National in the north and Conrail in the west. D. Proceedings Below On October 28, 1981, Guilford applied to the ICC for permission to acquire the Boston & Maine. Its plan to make the Boston & Maine profitable depended on three major elements: federal subsidy (in the form of low-interest loans); operating cost savings from merging the Boston & Maine with the Maine Central; and increased revenue from diverting to the Boston & Maine some of the east-west traffic that now travels over the Canadian National, the Canadian Pacific, or the Lamoille Valley. 1. Conditions Requested by the Boston & Maine’s Competitors Petitioner Lamoille Valley did not oppose the basic merger. Lamoille Valley claimed, however, that diversion of east-west traffic from it to the Boston & Maine would cause it to go bankrupt. Lamoille Valley therefore sought to obtain a substitute source of revenue by purchasing part of a Boston & Maine line that would carry some of the diverted traffic. Petitioner Canadian National also did not oppose the overall merger. It was concerned, however, that Guilford might seek to increase traffic diversion from the Maine Central/Canadian National east-west route to the Maine Central/Boston & Maine route by delaying the interchange of traffic between the Maine Central and the Canadian National. This would reduce the time advantage of the Maine Central/Canadian National route over the Maine Central/Boston & Maine route. Canadian National therefore asked the ICC to require Guilford to maintain current interchange service at Danville and Yarmouth Junctions, Maine. Petitioner Providence & Worcester is not a direct competitor of the Boston & Maine but rather carries north-south traffic that originates on the Boston & Maine to Connecticut and Rhode Island. Providence & Worcester opposed the merger on the grounds that the merged railroad would not be financially viable and would be unable to maintain current service over those Boston & Maine lines that connect with the Providence & Worcester. 2. The ICC’s Decision The ICC approved the merger and denied all requests for protective conditions. It found that the “quintessential public benefit” from the merger was “the emergence of the [Boston & Maine] from reorganization as part of a potentially strong unified rail network.” Boston & Maine Merger, 366 I.C.C. at 336. The Commission also found that “[t]he primary competitive impacts of the [merger] are favorable.” Id. at 340. There would be no direct reduction of competition because the Boston & Maine and the Maine Central “meet end-to-end” and “do not operate any parallel lines.” Id. at 341. In addition, the combined Maine Central/Boston & Maine would be a stronger competitor against “the Canadian lines and Conrail.” Id. . The ICC further found that the merger would not lead to cessation of essential services. The Commission rejected Canadian National’s request for protective conditions on traffic interchanges at Danville and Yarmouth Junctions for several reasons. In particular, it was not convinced that Guilford intended to delay interchanges. Also, Canadian National’s “ratemaking flexibility” gave it “sufficient competitive leverage to adequately combat any possible diversions of traffic.” Id. at 352. As for Lamoille Valley’s request to buy track from the Boston & Maine, the ICC agreed that the diversion of much of Lam-oille Valley’s overhead traffic to the Boston & Maine would be a “devastating” financial blow to Lamoille Valley. Id. at 353. Nevertheless, it rejected Lamoille Valley’s request because Lamoille Valley’s service was not “essential” — shippers could use truck transportation instead. While truck transportation was more expensive than rail transportation, it was not so expensive that shippers would be “forced out of business.” Id. E. Issues Presented Lamoille Valley, supported by intervenors Eastern Magnesia Talc Co. (a shipper that uses the Lamoille Valley) and State of Vermont, claims that the ICC’s test for whether a rail line provides “essential services”— will shippers go out of business if the line shuts down — is too strict and does not comport with the statutory directive that the ICC consider “adequacy of transportation to the public.” 49 U.S.C. § 11,344(b)(1)(A). Lamoille Valley also claims that the Corn-mission erred in not requiring Timothy Mellon, the sole owner of Guilford, to join the merger application and in failing to consider the interests of Lamoille Valley employees who may lose their jobs as a result of the merger. Canadian National joins Lamoille Valley in arguing that the ICC’s essential services test is too strict. It also charges that the Commission misanalyzed both Guilford’s incentive to delay interchanges at Danville and Yarmouth Junctions and the effect of such an action on Canadian National’s willingness to remain in competition with the Boston & Maine. Finally, Canadian National argues that the expedited procedural schedule used by the ICC in considering the merger was a “rule” within the meaning of the Administrative Procedure Act (APA), 5 U.S.C. §§ 551(4), 553, and was improperly issued without notice and comment. Providence & Worcester joins Lamoille Valley in objecting to the ICC’s failure to require Timothy Mellon to join the control application. In addition, it objects to the Commission’s finding that the Boston & Maine would be financially viable after the merger. Finally, Providence & Worcester raises several procedural objections: the ICC improperly failed to exercise jurisdiction over Guilford’s issuance of securities to Timothy Mellon in connection with the merger; the ICC erred in approving Guilford’s acquisition of the Vermont & Massachusetts Co. (which owns a single 56-mile long rail line leased by the Boston & Maine under a 999-year lease) without requiring a formal control application; and the ICC allowed Guilford prematurely to control the Boston & Maine. Intervenors Guilford and Trustees of the Boston & Maine support the ICC’s decision. In part II of this opinion, we defer, under principles of equitable abstention, to the reorganization court’s finding that the merged Boston & Maine will be financially viable, and conclude that the ICC properly approved the merger as a whole. In part III, we find that the ICC’s “essential services” test for determining when protective conditions may be needed, on its face, complies with the statute. Nevertheless, we find that the test, as applied to Lamoille Valley, is too strict. In part IV, we find that the ICC failed adequately to explain why Guilford would not downgrade service at Danville and Yarmouth Junctions and that the Commission placed undue weight on Canadian National’s system-wide revenues in assessing Canadian National’s willingness to provide service to Maine. In part V, we conclude that the ICC properly declined to take into account the interests of Lamoille Valley’s employees in continued employment. Part VI deals with the procedural issues. We conclude that the ICC’s failure to require Timothy Mellon to join the merger application, if error, is not prejudicial in this ease. We further conclude that the expedited procedural schedule used by the ICC is exempt from the APA’s notice and comment requirements as a “rule[ ] of agency ... procedure,” 5 U.S.C. § 553(b)(A), and that Providence & Worcester’s various procedural objections are without merit. Part VII is a conclusion. II. The Financial Viability of the Boston & Maine The ICC has a statutory responsibility to decide both whether the merger of the Maine Central with the Boston & Maine is “consistent with the public interest” under the Interstate Commerce Act, 49 U.S.C. § 11,344(c), and whether the reorganization of the Boston & Maine is “compatible with the public interest” under Bankruptcy Act § 77(d), 11 U.S.C. § 205(d) (1976) (current version at 11 U.S.C. § 1172). As part of its “public interest” determination under the Bankruptcy Act, the Commission found that the merged Boston & Maine would be financially viable. Boston & Maine Merger, 366 I.C.C. at 310-13. Finding that the other requirements of Bankruptcy Act § 77 were also satisfied, the Commission certified the reorganization plan to the reorganization court. Id. at 318. The reorganization court agreed that the merged Boston & Maine would be viable and approved the plan. In re Boston & Maine Corp., No. 70-250-M, mem. op. at 33-42 (D.Mass. opinion issued Feb. 23, 1983), appeal filed, No. 83-1086 (1st Cir. Feb. 3, 1983). In considering whether the merger was “consistent with the public interest” under Interstate Commerce Act § 11,344(c), the ICC saw “no need to repeat” its analysis of the viability of the reorganization plan. Relying on that analysis, the Commission again found that the merged Boston & Maine would be financially viable. Boston & Maine Merger, 366 I.C.C. at 339. Providence & Worcester claims that the ICC wrongly decided that the reorganized and merged Boston & Maine would be financially viable. The ICC and the Boston & Maine Trustees respond that we lack jurisdiction to review the Commission’s finding of viability because review of that finding lies exclusively in the reorganization court. We agree with this response, though we would characterize the issue as one of equitable abstention rather than lack of jurisdiction. The New Haven Inclusion Cases, 399 U.S. 392, 419-30, 90 S.Ct. 2054, 2072, 2077-2078, 26 L.Ed.2d 691 (1970), involved a situation similar to this one. The ICC had approved the merger of the bankrupt New Haven Railroad with the Penn Central Railroad. The parties appealed the Commission’s valuation of the New Haven’s assets both to a three-judge district court with jurisdiction to review the ICC’s approval of the merger and to the reorganization court charged with reviewing the New Haven’s reorganization plan. The two courts reached conflicting valuations. The Supreme Court held that the three-judge district court, although it had jurisdiction “in a technical sense,” id. at 427, 90 S.Ct. at 2076 should have “stayed its hand" and permitted the reorganization court to decide the valuation question, id. at 428, 90 S.Ct. at 2076-2077. We think abstention is likewise called for here. To be sure, in the New Haven Inclusion Cases, valuation was the sole issue still to be decided, so that the two courts were considering “identical issues.” Id. at 428, 90 S.Ct. at 2076-2077. Here, in contrast, we must decide numerous issues under the Interstate Commerce Act that were not and could not have been presented to the reorganization court. But the policy considerations that led the Court to order abstention in the New Haven Inclusion Cases also counsel abstention here. First, the standard for determining financial viability is the same under the Bankruptcy Act and the Interstate Commerce Act. Thus, there is “no danger that application of [Bankruptcy Act § 77] would yield results different from those to be produced by [Interstate Commerce Act § 11,344(c) ].” Id. at 424, 90 S.Ct. at 2075. Second, for this court to decide an issue that the reorganization court must also decide “ ‘would tend greatly to foment conflicts between coordinate courts and compel creditors, in the protection of their interests, to ride the circuit, demonstrating the basis of their positions in successive courts.’ ” Id. at 427, 90 S.Ct. at 2076 (quoting In re New York, N.H. & H.R. Co., 26 F.Supp. 18, 23 (D.Conn.), aff’d sub nom. Palmer v. Warren, 108 F.2d 164 (2d Cir.1939), aff’d, 310 U.S. 132, 60 S.Ct. 865, 84 L.Ed. 1118 (1940)). Third, Providence & Worcester has already had a full opportunity to present its arguments to the reorganization court, which rejected them. Thus, it cannot complain of unfairness because we do not give it a second bite at the apple. Cf. id. 399 U.S. at 428 n. 57, 90 S.Ct. at 2077 (“At no time has anyone questioned Penn Central’s status as a party litigant in the reorganization court or challenged its right to make a full presentation of its case there .... ”). Finally, the major policy reason counseling against abstention — “the importance of an expedited determination of the merits,” id. at 424, 90 S.Ct. at 2074 — is not a factor in this case since the reorganization court has already issued its decision. For the foregoing reasons, we defer to the reorganization court’s conclusion that the ICC properly found that the merged Boston & Maine would be financially viable. The parties raise no other objection to the substantive basis for the ICC’s approval of the merger as a whole, and we find in part VI infra no basis for reversing the Commission on procedural grounds. We therefore affirm the Commission’s decision that the merger is “consistent with the public interest.” 49 U.S.C. § 11.344(c). III. Lamoille Valley’s Request for Protective Conditions A. Standard of Review The standard of review for the ICC’s decision is well established. The Commission’s interpretation of its governing statute, like that of any regulatory agency, is “entitled to deference.” Federal Election Commission v. Democratic Senatorial Campaign Committee, 454 U.S. 27, 32, 102 S.Ct. 38, 42, 70 L.Ed.2d 23 (1981). We must accept the Commission’s interpretation if it is “sufficiently reasonable,” even if it is not “the only reasonable one or even the reading the court would have reached” on its own. Id. at 39,102 S.Ct. at 46; see National Wildlife Federation v. Gorsuch, 693 F.2d 156, 166-71 (D.C.Cir.1982). The ICC’s substantive conclusions are subject to review under the “substantial evidence” test of the Administrative Procedure Act, 5 U.S.C. § 706(2)(E). See Illinois Central Railroad v. Norfolk & Western Railway, 385 U.S. 57, 66, 87 S.Ct. 255, 260, 17 L.Ed.2d 162 (1966). Under that test, the Commission’s factual findings must be supported by enough evidence “to justify ... a refusal to direct a verdict” the other way in a case tried to a jury. Id. We also expect that when the facts are uncertain, the agency will “so state and go on to identify the considerations [it] found persuasive.” Industrial Union Department, AFL-CIO v. Hodgson, 499 F.2d 467, 476 (D.C.Cir.1974). Moreover, the Commission must “consider[ ] the relevant factors and articulate[] a rational connection between the facts found and the choice made.” Baltimore Gas & Electric Co. v. Natural Resources Defense Council, Inc., — U.S. —, —, 103 S.Ct. 2246, 2257, 76 L.Ed.2d 437 (1983); see Sierra Club v. Costle, 657 F.2d 298, 323 (D.C.Cir.1981). Having assured ourselves that the agency has fully considered an issue and given reasons for its conclusions, we have only limited power to second-guess those reasons.- “We can and do insist that the agency’s reasons and policy choices do ‘not deviate from or ignore the ascertainable legislative intent.’ ” Small Refiner Lead Phase-Down Task Force v. United States Environmental Protection Agency, 705 F.2d 506, 520 (D.C.Cir.1983) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841, 850 (D.C.Cir.1970), cert, denied, 403 U.S. 923, 91 S.Ct. 2233, 29 L.Ed.2d 701 (1971)). Beyond that, however, “[i]t is not for the court to strike down conclusions that are reasonably drawn from the evidence and findings in the case.” Illinois Central, 385 U.S. at 69, 87 S.Ct. at 262; see Seaboard Coast Line Railroad v. United States, 599 F.2d 650, 652 (5th Cir.1979). B. The ICC’s Decision The Lamoille Valley operates a single 98-mile-long east-west line between St. Johnsbury, Vt. in the east (where it connects with the Maine Central) and Swanton, Vt. in the west (where it connects with the Canadian National). It depends for financial viability on “overhead” traffic which it carries from the Maine Central to the Canadian National or vice-versa. The merger of the Maine Central with the Boston & Maine gives Guilford an incentive to divert traffic from the Maine Central/Lamoille Valley/Canadian National line to the competing Maine Central/Boston & Maine line. Guilford estimated that its diversion efforts would reduce Lamoille Valley’s gross revenues by about $180,000, based on 1980 traffic data. See Boston & Maine Merger app. C, 366 I.C.C. at 370. Lamoille Valley challenged Guilford’s assumptions and counterestimated traffic diversion of as much as $300,000. The ICC found flaws in Guilford’s analysis and concluded that Guilford’s diversion estimates — both for the Lamoille Valley and for other carriers— “are understated and can be considered as bare minimum estimates of traffic and revenue diversions.” Id. at 372. For the Lamoille Valley, the ICC found it unnecessary to make its own estimate of traffic diversion. It concluded that “[e]ven using [Guilford’s] figures, which we have found to be minimum estimates, [Lamoille Valley] stands to lose .. . over 15 percent of its 1980 gross revenue. Such a financial blow would be devastating to [Lamoille Valley].’’ Id. at 353. Despite this bleak outlook, the ICC denied Lamoille Valley’s request for protec-five conditions because Lamoille Valley’s services were not “essential.” Lamoille Valley’s overhead service was not essential “since a number of other connections to the Canadian routes exist.” Id. at 354. Moreover, trucks could substitute for Lamoille Valley’s local service. While truck service would undeniably cost more, shippers sometimes used trucks even now and no shippers would be “forced out of business” if they had to rely exclusively on trucks: Even assuming that ... there would be a cessation of operations, we cannot conclude that essential services would be affected. It appears from the evidence presented by shippers ... that they can be served by motor carriers. All three supporting shippers ... are either presently being served in part, or have been served previously, by truck.... While motor carrier service is more costly, it is a reasonably adequate substitute for [rail] service. None of the supporting shippers has indicated that it would be forced out of business if [Lamoille Valley] discontinued service. Id. at 353. Relying on this passage, Lamoille Valley asserts that the Commission used a “business termination” test to determine if Lam-oille Valley’s local service was “essential,” and that this test does not comply with the statutory directive that the ICC consider “adequacy” of transportation. 49 U.S.C. § 11,344(b)(1)(A). Lamoille Valley also claims that the ICC failed to analyze the reduction in competition for east-west traffic from the Lamoille Valley going out of business. ICC counsel replies that the Commission, in finding that Lamoille Valley’s services were not “essential,” relied solely on the fact that the shippers were now or recently had been served by truck. Counsel asserts that the Commission has “never adopted or recommended” a business termination test and that the Commission’s reference to shippers being “forced out of business” was mere dictum. As for possible reduction in competition from the Lamoille Valley going out of business, the Commission evaluated it and found it to be negligible. C. The Essential Services Test The ICC’s policy statement on railroad mergers explains that there are “two potential results from consolidations which would ill serve the public — reduction of competition and harm to essential services.” 49 C.F.R. § 1180.1(c)(2) (1982). The Commission defines “essential services” as: A service is essential if there is a sufficient public need for the service and adequate alternative transportation is not available. Id. § 1180.1(c)(2)(ii). The ICC treats the existence of harm to competition or to essential services as a threshold test for when conditions may be needed to reduce the adverse effects of a merger. It will never impose protective conditions unless a merger will cause significant reduction in competition or harm to essential services. If a merger will reduce competition or harm essential services, the Commission will decide whether conditions are desirable by balancing the costs and benefits of imposing conditions. This subsection considers the ICC’s use of conditions to protect essential services; subsection D considers the ICC’s use of conditions to preserve competition. In general, the ICC’s use of the “essential services” test as a threshold test for when protective conditions deserve further consideration is unobjectionable. Leaving aside the requirement — not at issue in this case — of a. “sufficient public need,” the ICC defines an “essential service” as one for which an “adequate alternative ... is not available.” This definition simply restates the statutory requirement that the ICC consider “adequacy of transportation to the public.” 49 U.S.C. § 11,344(b)(1)(A). Accord Brotherhood of Maintenance of Way Employees v. ICC, 698 F.2d 315, 319 n. 14 (7th Cir.1983) (“Essential services are basically those the deprivation of which would result in inadequate service to the public.”); of. Missouri-Kansas-Texas Railroad v. United States, 632 F.2d 392, 400-04 (5th Cir. 1980) (approving the ICC’s use of an earlier definition of “essential services”), cert, denied, 451 U.S. 1017, 101 S.Ct. 3004, 69 L.Ed.2d 388 (1981). It remains to consider whether the ICC properly applied the essential services test to Lamoille Valley. The ICC gave three reasons for concluding that there were adequate alternatives to Lamoille Valley’s rail service. First, all three shippers who testified in favor of Lamoille Valley’s request for protective conditions “are either presently being served in part, or have been served previously, by truck.” Furthermore, one shipper, E.T. & H.K. Ide Co., “has access to [Canadian Pacific] rail in St. Johns-bury, [Vt.]” Finally, no shipper claimed that “it would be forced out of business if [Lamoille Valley] discontinued service.” Boston & Maine Merger, 366 I.C.C. at 353. 1. Current or Past Truck Service The Commission’s first reason for believing that Lamoille Valley’s service is not essential has some probative force. It is plausible that a shipper formerly served by truck can again use trucks and that a shipper now served partially by truck can adequately be served entirely by truck. We must still determine, however, whether the record supports the Commission’s belief that the individual shippers now served by the Lamoille Valley can be adequately served by truck. We pass quickly over the one shipper, E.T. & H.K. Ide Co., which has access to Canadian Pacific’s rail service, for that service is surely an adequate substitute for Lamoille Valley’s service. For a second shipper, Lamoille Grain, the record shows that the price difference between rail and truck service is small. Indeed, as recently as 1980, Lamoille Grain switched from rail to truck because truck rates were slightly lower. It switched back only after Lamoille Valley reduced its rates to make them competitive with truck rates. Thus, Lamoille Grain can be served adequately by truck. The third shipper, Eastern Magnesia Talc Co., presents a much harder case. It operates a talc mine from which it ships talc by truck to some nearby locations and by rail (usually 100-ton hopper cars) to more distant locations. Eastern Magnesia Talc explained to the ICC that the cost of “loading to truck, hauling to a [railhead], unloading, ... and loading to railroad cars” would be “prohibitively expensive” and would effectively prevent it from serving distant locations. In light of Eastern Magnesia Talc’s un-contradicted statement about the need for rail service for a substantial part of its business, the Commission could not reasonably conclude that because Eastern Magnesia Talc uses trucks for part of its business, trucks are an adequate substitute for rail for the remainder of its business. We turn therefore to the Commission’s alternate reason for concluding that Eastern Magnesia Talc could switch to truck service — that Eastern Magnesia Talc had not asserted that it would be “forced out of business” if it loses rail service. 2. Business Termination As an initial matter, and despite Commission counsel’s denial, it seems clear to us that the ICC is indeed using a “business termination” test to determine whether a substitute for rail service is “adequate.” The ICC’s explanation, though terse, is unambiguous: While motor carrier service is more costly, it is a reasonably adequate substitute for [rail] service. None of the supporting shippers has indicated that it would be forced out of business if [Lamoille Valley] discontinued service. Boston & Maine Merger, 366 I.C.C. at 353 (footnote omitted). Similarly, in approving Guilford’s acquisition of the Delaware & Hudson, the ICC explained that the north-south service provided by Canadian National was not essential because, in part: Nowhere in the record has it been shown that any of [its] customers would go out of business if [Canadian National’s] rail service would be terminated. Delaware & Hudson Merger, supra note 1, 366 I.C.C. at 420. The ICC has used similar phraseology in several recent cases approving abandonment of rail service. See, e.g., Missouri Pacific Railroad — Abandonment— Between Port Barre & Jefferson Island, Docket No. AB-3 (Sub-No. 27), slip op. at 5 (July 21, 1982) (not printed) (protesting shipper did “not contend that the loss of rail service would put it out of business”). We can find no justification in the statute for use of a “business termination” test to determine whether alternative service is adequate, and indeed neither the ICC nor Guilford offers any. The legislative history does not explain what Congress meant when it instructed the ICC to consider “adequacy of transportation to the public.” Thus, we must give the word “adequate” its ordinary meaning. A refusal to consider protective conditions unless alternative service is so expensive that shippers will be forced out of business does not comport with the statutory directive to take into account “adequacy” of transportation. The term “adequate,” as the Supreme Court noted long ago in a similar context, is “a relative expression.” Atlantic Coast Line Railroad v. Wharton, 207 U.S. 328, 335 (1907). For a shipper, loss of a service that it currently uses generally portends some decrease in profit, which can be anywhere from minor to devastating. If the additional cost involved in using an alternative service is small, the ICC need not be concerned. At the other extreme, if the added cost is so grave as to force bankruptcy, the alternate service is clearly inadequate, at least for that shipper. Somewhere between these two extremes, we pass from adequate to inadequate service. Within reason, the Commission has discretion to draw the dividing line as it sees fit. The business termination test, however, is too far to one extreme to be a reasonable definition of “adequacy.” In seeking on remand to draw a better line, the Commission should inquire into how much more costly a proposed alternative is and whether loss of existing service will cause substantial harm to the local economy or to shippers who now use that service. With regard to individual shippers, if a shipper can use alternate service and still earn a fair return on capital, sufficient to generally maintain current operations and justify reinvestment of earnings, the alternate service would seem adequate. The same is true if the shipper can switch production to another plant without serious harm to the local economy. On the other hand, if a shipper cannot earn a fair return, or can do so only by sharply curtailing operations, the Commission probably ought to inquire further into the desirability of protective conditions. It is worth stressing that the essential services test is not a test for imposing conditions, but a test for when it is worthwhile to consider doing so. The ICC may do well, in close cases, to err on the side of reaching the merits of the underlying question whether protective conditions are in the public interest. Sometimes, the profitability of individual shippers may be hard to assess, or a railroad may serve a large number of small shippers, insignificant individually but important in aggregate. The Commission may then want to consider whether most shippers in a particular industry rely on rail service. If not, that suggests that alternate transportation is probably an adequate substitute for shippers in that industry. If most shippers do rely on rail service, then alternate service is probably not an adequate substitute. The Commission will also need to assess the importance of individual shippers. If the major shipper or shippers on a rail line can switch to truck, the Commission can reasonably find that truck service is adequate even if a few minor shippers cannot switch. Conversely, alternate service may be inadequate if a major shipper whose facility is important to the local economy cannot switch, even if most shippers can use the alternate service. Again, we urge the Commission, in cases of doubt, to proceed beyond the threshold “essential services” stage of its inquiry. We note that our interpretation of “adequacy” of alternative transportation as meaning more than just not going out of business is consistent with past ICC interpretations. While there is no useful case law involving protective conditions for mergers (until recently, the ICC imposed a set of standard conditions on virtually every rail merger), abandonment cases provide a close analogy. The abandonment cases do not define “adequacy,” but they do show that the ICC, in determining the public need for rail service, has inquired into the relative costs of rail and truck service. For example, in Northwestern Pacific Railroad Abandonment (Portion) Sausalito Branch, 312 I.C.C. 783, 789 (1962), the Commission found that abandonment was not warranted even though most of the shippers objecting to the abandonment “use motortrucks for part of their shipments.” The Commission noted that a distiller “would be handicapped by the narrowing of the area in which it could purchase raw materials”; a dealer in heavy diesel engines would incur “added expense” in shipping disassembled engines by truck and assembling them at destination; and a sailboat manufacturer would incur “40 to 50 percent higher [costs]” in shipping sailboats by truck. Id. The Commission did not insist that the affected shippers claim they would be forced out of business by loss of rail service. Recent court decisions are in accord. See Georgia Public Service Commission v. United States, 704 F.2d 538, 545 (11th Cir.1983) (reversing ICC approval of a rail abandonment): The uncontroverted testimony ... is that ... the “additional” transportation costs [of truck service] would be prohibitive .... If the phrase “alternative [transportation]” is to have any meaning it must be interpreted to include transportation both logistically and economically feasible. See also Indiana Sugars, Inc. v. ICC, 694 F.2d 1098,1101 (7th Cir.1982) (abandonment would “inflict serious hardships” on sugar company that depended on rail for much of its inbound traffic). 3. Conclusion We conclude that the ICC misconstrued its governing statute and failed to justify its belief that Eastern Magnesia Talc can be served adequately by trucks. Therefore, we remand to the ICC to reconsider Eastern Magnesia Talc’s ability to switch to trucks. If Eastern Magnesia Talc cannot switch, the Commission will also need to consider two questions this opinion does not address — whether the needs of a single shipper make Lamoille Valley’s service essential and, if so, whether the protective conditions sought by Lamoille Valley are in the public interest. We order as follows: As discussed in part II supra, the merger as a whole is approved, but the ICC shall expeditiously consider whether it is in the public interest to impose protective conditions in order to preserve the Lamoille Valley’s service. D. Anticompetitive Effect In reviewing a merger, the ICC must consider not only adequacy of transportation but also “whether the proposed transaction would have an adverse effect on competition among rail carriers in the affected region.” 49 U.S.C. § 11,344(b)(1)(E). The ICC found that the Maine Central and the Boston & Maine “do not operate any parallel lines, but meet end-to-end.” Boston & Maine Merger, 366 I.C.C. at 341. Therefore, the merger would not “directly reduce” competition, except for a minor impact on shippers in Portland, Maine (where the two rail systems meet). Id. The Commission found that, on the contrary, by strengthening the Boston & Maine, the merger would enhance competition for east-west traffic between the Boston & Maine route and the currently dominant Canadian National and Canadian Pacific routes. Id. at 340-41. The Commission recognized that “[a]n end-to-end consolidation may allow the [combined] system to divert traffic from remaining competitors and thereby foreclose their opportunity to compete in the marketplace.” Id. at 341 n. 35. It concluded, however, that the possible demise of the Lamoille Valley would have “negligible” competitive impact because “a number of other connections to the Canadian routes exist.” Id. at 354. Lamoille Valley asserts that the Maine Central does in fact compete with the Boston & Maine for east-west traffic. We agree. Traffic from Maine can travel west over four routes: (1) over the Canadian National (which connects with the Maine Central at Danville and Yarmouth Junctions, Maine) to Montreal and points west; (2) over the Maine Central’s Mountain Division line to St. Johnsbury, Vt., and from there over the Lamoille Valley to the Canadian National to Montreal and points west; (3) over the Maine Central’s Mountain Division line to St. Johnsbury and from there over the Canadian Pacific to Montreal and points west; and (4) over the Boston & Maine to Albany and points west. Thus, the Maine Central’s Mountain Division line competes with the Boston & Maine’s east-west line. At present, the Maine Central has an incentive to route traffic over its own long haul (routes 2 and 3), in preference to the Canadian National route or the Boston & Maine route. After the merger, Guilford will prefer to route traffic to its new long haul — the Boston & Maine route — in preference to the other three routes. This traffic shift will lead to the Boston & Maine gaining revenue ($3.2 million annually according to Guilford’s traffic study) while the Maine Central loses revenue ($1.9 million annually according to Guilford’s study) due to decreased traffic over its Mountain Division line. See Boston & Maine Merger app. C, 366 I.C.C. at 370. We must therefore reject the ICC’s statement that the Maine Central and the Boston & Maine do not compete with each other. Nevertheless, we believe that the record supports the ICC’s belief that the “primary competitive impacts of the [merger] are favorable.” Id. at 340. The east-west route from Maine to the Midwest is over 1000 miles long. Currently, the Canadian National and the Canadian Pacific are the dominant carriers. The Boston & Maine’s more circuitous route provides minimal competition, especially given the Boston & Maine’s sorry finances and the Maine Central’s incentive to route traffic away from the Boston & Maine and onto its own Mountain Division line and thence to the Canadian carriers. The merger, the ICC found, will permit the Boston & Maine to compete effectively over a large portion of this route. For that benefit, decreased competition over the short distance traversed by the Maine Central’s Mountain Division line is a small price to pay. As for reduction in competition due to the demise of the Lamoille Valley, we note that a large number of competing carriers on a particular route is not an end in and of itself. Rather, competition is valuable because it promotes efficient service. We find reasonable the ICC’s belief that the remaining competition among the Canadian National, the Canadian Pacific, and the Boston & Maine will provide sufficient stimulus for efficient operation of rail lines. IV. Canadian National’s Request for Protective Conditions A. The ICC's Decision Canadian National carries freight from Danville and Yarmouth Junctions, Maine (where it connects with the Maine Central) west to Montreal and Chicago. The joint Maine Central/Canadian National line is the fastest route between Maine and Chicago. The Canadian National’s time advantage depends on close coordination of schedules and speedy interchange with the Maine Central at Danville and Yarmouth Junctions. Canadian National expressed concern to the ICC that after the merger, Guilford would refuse to cooperate with Canadian National on scheduling and interchanges, thus increasing transit time over the Maine Central/Canadian National route. This would lessen the attractiveness of that route and permit Guilford to divert more traffic to the currently slower Boston & Maine route. Canadian National also claimed that reduced traffic would force it to cut expenses by running fewer trains and reducing track maintenance, which would further increase transit time, which would lead to additional diversion of traffic to the Boston & Maine, which would force Canadian National to reduce service further, and so on in a vicious cycle that ultimately would lead Canadian National to abandon its Grand Trunk Eastern line between Maine and Montreal. Guilford’s traffic study, which assumed no increase in interchange time at Danville and Yarmouth Junctions, predicted that diversion of east-west traffic to the Boston & Maine would cost Canadian National $2 million in annual revenues. Canadian National counterestimated diversion losses of $8.1 million per year if Guilford did not downgrade the Danville and Yarmouth interchanges and $12.5 million if Guilford did downgrade the interchanges. Canadian National asked the ICC to require Guilford to maintain “present operating arrangements, schedules of service, [and] blocking and handling of freight traffic interchanged at [Danville and Yar-mouth] junctions.” Boston & Maine Merger app. B, 366 I.C.C. at 365. It claimed that fast service to Chicago is “essential” to shippers and that these conditions were needed both to preserve that service and to preserve competition for east-west traffic between it and the Boston & Maine. In response, the ICC gave several reasons for believing that Canadian National’s service would not be adversely affected by the merger. It did not reach the question whether that service is essential, nor did it discuss whether loss of the Canadian National’s Grand Trunk Eastern line would significantly reduce competition for east-west traffic. First, the ICC explained that Canadian National would not have to abandon its Grand Trunk Eastern line because even Canadian National’s “worst-case scenario” of $12.5 million diversion “represent[s] a mere fraction of 1 percent of CN’s 1980 gross freight revenue.” Id. at 352. Second, there was “no indication in the record that [Guilford] has any intention of downgrading this interchange relationship.” Id. Third, Guilford would not have an incentive to downgrade “since substantial traffic through these interchanges is nondiverti-ble.” Id. Finally, Canadian National “possesses sufficient competitive leverage to adequately combat any possible diversions of traffic” by virtue of its “single-line system between [Maine and the midwest] and the ratemaking flexibility inherent in such a system.” Id. We have substantial difficulties with each of these arguments. Cumulatively, the weaknesses in the ICC’s arguments lead us to remand to the Commission to reconsider whether Canadian National’s service will be adversely affected by the merger and, if so, whether the public interest would be advanced by restricting Guilford’s ability to downgrade its interchanges with Canadian National. B. The Limited Relevance of Systemwide Revenues We consider first the ICC’s belief that essential services would not be affected because the diversions from the Canadian National to the Boston & Maine represent a fraction of 1% of Canadian National’s gross revenues and thus, presumably, Canadian National could underwrite any losses on its Grand Trunk Eastern line. We fail to see the relevance of Canadian National’s system-wide revenues. First, if Guilford downgrades interchanges at Danville and Yarmouth Junctions, shippers will lose the benefits of “the best transit time between Maine and the Midwest.” Boston & Maine Merger, 366 I.C.C. at 352. Canadian National’s ability to absorb the resulting revenue loss will not change that fact. Second, the ICC’s emphasis on Canadian National’s financial strength and neglect of the profitability of the Grand Trunk Eastern line ignores Canadian National’s incentive to maximize profits by shedding money-losing operations. Nothing in the record supports the implausible proposition that Canadian National will indefinitely subsidize losses on its Grand Trunk Eastern line. On the contrary, undisputed testimony from Canadian National’s President establishes that Canadian National treats its branches and subsidiaries as independent “profit centers” and will not subsidize a losing branch line. Third, the ICC itself, as a policy matter, opposes on efficiency grounds the subsidizing of money-losing lines. For example, the Commission will not “under any circumstances” require one carrier to indemnify another carrier for losses due to traffic diversion, because “[t]his transfer of funds would, in effect, be a massive cross-subsidization ... of inefficient operations of another railroad.” Burlington Northern— Control — St. Louis-S.F., supra note 37, 360 I.C.C. at 952-53. Similarly, in abandonment proceedings, the Commission analyzes primarily the profitability of the line sought to be abandoned and gives little weight to the profitability of the parent railroad. See, eg., Illinois Central Gulf Railroad Abandonment Between Herscher & Barnes, 363 I.C.C. 690, 703 (1980) (“Unprofitable operations ... burden [the carrier] and interstate commerce ... [and the carrier] should not be required to continue [such] operations ..., regardless of whether it is financially capable of doing so .... ”), aff’d sub nom. Bloomer Shippers Association v. ICC, 679 F.2d 668 (7th Cir.1982). The ICC, to be true to this policy, should not expect Canadian National to operate an inefficient, money-losing line. In sum, the ICC’s belief that Canadian National’s overall profitability would permit Canadian National to preserve existing service on its Grand Trunk Eastern line ignores Guilford’s unilateral ability to downgrade interchange service, is unsupported by evidence in the record, and is contrary to Commission policy. This is not to say that Canadian National’s Grand Trunk Eastern line will be unprofitable once the Maine Central merges with the Boston & Maine. Canadian National asserts that the Grand Trunk Eastern line will sustain heavy losses as a result of diversion of traffic to the Boston & Maine, but Guilford disputes this. On remand, the ICC will have an opportunity to decide this factual question. C. Guilford’s Intent to Maintain Present Service A second major factor leading the ICC to conclude that the merger would not affect Canadian National’s service was that the record gave “no indication” that Guilford intended to downgrade service at the Dan-ville and Yarmouth interchanges. 366 I.C.C. at 352. We think this factor is entitled to little if any weight. First, the record shows only that Guilford has no present intent to downgrade interchange service. Guilford made no promises concerning the future. Thus, Guilford’s owner, Mr. Mellon, when asked if he would “give Guilford’s assurance” that it would preserve service at Danville and Yarmouth Junctions, gave an equivocal answer — “[i]f the economic circumstances are somewhat what they are today.” Moreover, Guil-ford opposed Canadian National’s request for conditions, thus reserving for itself the right to decrease service in the future. Second, as the Commission itself has recently recognized, self-serving statements by a merging railroad’s officers are entitled to little credence. See Norfolk & Western — Control—Detroit, Toledo & Iron-ton, supra note 27, 360 I.C.C. at 512 (promise by two competing railroads to operate independently a third competing line is “self-serving and must be examined closely .... It is difficult to expect sacrificial behavior from profit seeking corporations into the indefinite future.”). Guilford will, we expect, seek to maximize profits. If it can profitably downgrade interchanges with Canadian National, Guilford will have strong incentives to do so, notwithstanding its current disclaimers. Thus, Guilford’s actual financial incentives, not its professed intent, must be the dominant concern. We consider those incentives next. D. Guilford’s Incentives to Downgrade Interchange Service The ICC’s third reason for rejecting protective conditions at Danville and Yar-mouth Junctions was that “it would not appear to be to [Guilford’s] benefit for such deterioration [of interchange service] to take place, since substantial traffic through those interchanges is nondivertable.” Boston & Maine Merger, 366 I.C.C. at 352. We think that the Commission failed to analyze properly Guilford’s incentives to downgrade interchange service. 1. Guilford’s Incentives in Antitrust Theory The Maine Central is the only railroad providing east-west service to the Maine paper industry. Moreover, Canadian National claims, and the Commission does not dispute, that most of the Maine Central’s traffic is captive to rail — trucks are not a significant source of competition. Thus, the Maine Central has market power over rail traffic to and from Maine. As a theoretical matter, then, the Maine Central may be able to increase profits by altering its interchange service. The merger of the Maine Central with the Boston & Maine is, in antitrust terminology, a “vertical” merger. Ordinarily, a vertically integrated monopolist has no incentive to use its monopoly power over one level of production (rail service in Maine) to increase profits at another level (rail service from Maine to the midwest). As the leading treatise puts it, “there is but one maximum monopoly profit to be gained” from a monopoly of one level of production, and that profit may be gained directly at the monopolized level (here, rail service in Maine) through appropriate pricing. 3 P. Areeda & D. Turner, Antitrust Law § 725b (1978). There is an exception to this rule, however, for a regulated monopolist, which may be able to obtain from a second level of production “the monopoly profits which effective regulation of the franchised monopoly precludes.” Id. ¶ 726e (footnote omitted). The Maine Central is, of course, a regulated monopolist. It is barred by the ICC from charging a monopoly price for rail service to Maine, but may be able to use its control over service at Danville and Yar-mouth Junctions to shift its potential monopoly profit to the Boston & Maine. Indeed, the ICC itself has recently recognized the risk of such conduct. See Union Pacific — Control—Missouri Pacific, supra note 21, 366 I.C.C. at 529: [In an end-to-end merger], the consolidated system may profit by limiting the “downstream” competition [from connecting carriers] and routing traffic over its own lines from origin to destination. ... This vertical foreclosure effect will occur, if at all, at the gateways served commonly by the conso