dissenting.
I regret that I am unable to join in Judge Kennedy’s thoughtful, comprehensive majority opinion. I am unable to do so because in my view the ICC has erroneously interpreted and applied the Panama Canal Act, 49 U.S.C. § 11321 (1982) (the “Act”), so as to approve for the first time during the seventy-three year history of the Act the acquisition by a railroad of a barge-line.1 Today’s decision by our Court is one of first impression among the courts of the United States, for no court has ever approved such an acquisition. The acquisition in question is one of enormous importance: it concerns the acquisition by the nation’s second largest railroad (CSX Corporation) of American Commercial Lines, Inc. (ACL), which owns the nation’s largest bargeline. In approving the acquisition, the ICC rejected and overruled its own prior controlling decisions.
While the independent regulatory agencies, such as the ICC, have broad discretionary power when acting pursuant to their statutory mandates, that discretion is not unbridled. It is the responsibility of the reviewing courts to see to it that the law has been properly interpreted and applied, and that the decision in question has been based on substantial evidence.2 Northern Lines Merger Cases, 396 U.S. 491, 503 (1970). Regardless of the ultimate merits, or even lawfulness, of the acquisition here in question, I am convinced that the ICC plainly failed to apply the proper legal standards under the Act. In short, the ICC’s new interpretation of the Act is not “sufficiently reasonable”. See majority op. at 1183. In my view, the ICC decision should be vacated and the case should be remanded for proceedings in accordance with proper legal standards. See Coal Exporters Ass’n v. United States, 745 F.2d 76, 80 (D.C.Cir.1984) cert denied, 105 S.Ct. 2151 (1985).
I.
Several aspects of the ICC’s decision raise especially serious doubts about the result reached. I agree with the majority’s holding that the Act does not provide for an absolute prohibition of railroad ownership of bargelines. I do take issue, however, with the method by which the ICC and the majority conclude that this acquisition is permitted under the Act. The essential, and fatal, error upon which the ICC’s decision is based, is its utter failure to consider the effect of this acquisition on competition between railroads and barge-lines. Since the very purpose of the Act was to foster and maintain rail/barge competition, this failure strikes me as curious. Indeed, it suggests an attempt to frame the analysis in such a way as to support a result-oriented decision.
*1197While I acknowledge that the majority’s statement of the facts is straightforward and accurate, the following brief reference to some of the relevant facts as found by the ICC may help to place in perspective the issue to which this dissent is addressed.
CSX and its rail subsidiaries operate over 27,000 miles of track. CSX is the leading carrier of solid bulk commodities by any mode. In 1981, for example, CSX moved 541 million tons of traffic. This constituted one-half of all coal and chemical shipments and nearly one-third of all agricultural products shipped by rail in the eastern United States. Most importantly, CSX controls primary access to one-half of all the rail/barge coal-loading capacity on the Ohio River system, including all of the terminals along one important 300 mile stretch of the Ohio River. CSX and two other railroads account for 86% of all eastern Class I rail revenues. Revenues for 1982 totalled nearly $5 billion, with net income for that year amounting to $338,-400,000.
ACL operates its barge lines over 7,500 miles of the 15,000 miles of inland waterways. ACL is the nation’s largest line-haul bargeline. ACL is one of the top three for-hire bargelines in the transport of coal, chemicals and agricultural products on the inland waterways. ACL and five other carriers transport 70% of all eastern coal shipments to utilities along water routes. ACL’s revenues for 1982 totalled $295,000,-000, with net income for that year amounting to $21,000,000.
CSX and ACL directly compete for customers to transport coal, chemicals and agricultural products in 12 states and 59 metropolitan areas, including Chicago, Cincinnati, Louisville, Memphis, Mobile, New Orleans and Pittsburgh. CSX and ACL directly compete at 163 points along the inland waterway system in these states. Thirty utility plants have the capacity to receive either the rail or water services provided by CSX and ACL.
II.
The ICC found that CSX and ACL compete in three ways. First, they directly compete for business to and from the 163 points along the water route mentioned above and to and from points within a band extending 25 to 200 miles on either side of the water route, depending on circumstances. Second, they compete for-service to the same destination from different points of origin where the end-user has the option of buying from two different sellers, one with access to water routes and the other with access to rail routes. Third, there is some competition for the carriage of goods and commodities where both destination and origin are different but the products compete in the same end-use market.
Thus, the ICC found that ACL and CSX actually do compete for traffic. Indeed, absent such direct competition, ICC approval of the acquisition would not be required. The Act only prohibits a rail carrier from maintaining an interest in a water carrier “with which it does or may compete for traffic.” 49 U.S.C. § 11321(a)(1) (1982). The ICC referred to specific examples of instances in which CSX and ACL sought traffic from the same shipper, including a bidding competition to supply the Tennessee Valley Authority’s Cumberland, Tennessee plant. CSX and ACL submitted the two lowest bids, with ACL eventually winning the contract to supply 120 million tons of coal over a 20 year period.
Nevertheless, in pursuing its analysis under § 11321(b), the ICC completely disregarded the effect of the acquisition on rail/barge competition and considered only competition among water carriers. In defining the relevant product market, the ICC stated it would “focus on water carrier transportation”. (Dec. at 24). In its view, railroads are not effective constraints on barge rates because barge rates are so much lower. I believe that this conclusion reached by the ICC is highly questionable and probably erroneous — for the following reasons.
First, this conclusion ignores the ICC’s other finding that ACL and CSX directly compete for the same traffic. Second, it *1198ignores specific instances in which CSX and ACL were the two lowest bidders on the same contract, such as the TVA’s Cumberland plant contract. Third, it ignores the historic ability of railroads in general, and CSX in particular (as demonstrated by the evidence), to reduce rates to levels that are competitive with barges. P.D. Locklin, Economics of Transportation 729 (6th ed. 1966); Mapes, Competition Between Railroads and Water Carriers: A Comparison of the Regulatory and Antitrust Approaches and a Proposal for Reform, 39 U.Pitt.L.Rev. 653, 657 (1978). Fourth, it ignores the ICC’s own prior decisions in which it relied on the competition provided by barges to justify railroad mergers. In CSX Corp. — Control—Chessie System, Inc., and Seaboard Coast Line Indus., Inc., 363 I.C.C. 521 (1980), for example, the ICC approved the consolidation of three railroads which now constitute the present CSX. The ICC stated at that time that “[t]he affiliated carriers will be able to compete more effectively with truck and barge operations” (emphasis added) by offering “single-system service” to shippers. Id. at 563. See also Norfolk Southern Corp. — Control—Norfolk & W. Ry., 366 I.C.C. 173, 201-02 (1983). The ICC’s convenient manipulation of the facts and the law to reach the result it did in the instant case strikes me as nothing short of astounding.
The ICC’s sleight of hand, moreover, does not end there. On this appeal, it argues that § 11321(b) is concerned only with competition among water carriers on the route in question and that competition between the rail and water carriers involved in the acquisition is irrelevant. It argues that § 11321(b) would be superfluous if rail/barge competition were considered because § 11321(a) requires direct competition for the section to be operable and in every such case some reduction of competition will result.
In its decision in this very case, however, the ICC has defined the competitive analysis under § 11321(b) as follows:
“The involved competition has many facets. Competition between the involved water carrier and the acquiring railroads, between the involved water carrier and other water carriers, and among the non-included water carriers all may effect the level of competition on the water route in question.” (Dec. at 18) (emphasis added)
This standard is consistent with the ICC’s decisions in prior similar cases. E.g., Illinois Central R. Co.-Control-John I. Hay Co., 317 I.C.C. 39, 55 (1962); Lake Line Applications Under Panama Canal Act, 33 I.C.C. 699, 715-16 (1915). In short, the ICC either has failed to follow its own interpretation of the statute in its decision in the instant case or it seeks to alter that interpretation on appeal in an attempt to avoid what even it thought would likely be an adverse result.
Unhappily, this is not the first time that the ICC has resorted to such questionable attempts to impose its will over that of Congress. It has not been successful in the past. As the District of Columbia Circuit so succinctly stated in vacating another result-oriented ICC decision, “[a]n agency cannot hide the standards under which it operates, for we are unable to evaluate whether its reasoning meets the reasoned decision making requirement unless we know against what standards its factual findings have been judged.” Coal Exporters, supra, 745 F.2d at 99; see also id. at 90.
Whatever deference must be accorded “the interpretation of a statute by an agency charged with its enforcement”, Meade Township v. Andrus, 695 F.2d 1006, 1009 (6th Cir.1982), such an agency, at the very least, must be held to its own interpretation. While an agency may not be “disqualified from changing its mind”, it must give sufficient reasons for its new interpretation. Otherwise, a reviewing court cannot “approach the statutory construction issue ... with[ ] regard to the administrative understanding of the statute[ ]." NLRB v. Local Union No. 103, 434 U.S. 335, 351(1978); see majority op. at 1186. No reasons whatsoever were given by the ICC *1199for the broad change in its interpretation of the Act in the instant case.
Moreover, the ICC’s new interpretation of the Act utterly fails to satisfy Congress’s intent to prohibit all rail/barge mergers that effect any reduction in competition on the water routes in question. 49 U.S.C. § 11321(b) (ICC may permit merger of rail and bargelines only if “it will still allow competition, without reduction, on the water route in question.”) (emphasis added). Aside from the ICC’s own past practice of examining the level of rail/barge competition, basic competition principles require such an examination to determine the effect of the merger, not only on rail/barge competition, but on competition among bargelines as well.
This merger has vertical as well as horizontal aspects in that CSX transports cargo from inland points to points on the water’s edge for loading on barges. Of all the coal shippers in CSX’s operating region, 85% are limited to using CSX. Compounding this control is CSX’s substantial control of rail/barge loading terminals along the most significant coal transport water routes. Accordingly, the ICC concluded that “the merger is likely to lead to diversion to [ACL] from other carriers of some CSX-originated traffic”. (Dee. at 44). The ICC assumes, however, that only a small amount of such traffic will be diverted to ACL because ACL will not always be the most “cost effective barge partner for CSX”. (Dec. at 43).
Even if only “some” traffic in this concentrated market is diverted to ACL solely because ACL is affiliated with CSX, necessarily there has been a prohibited reduction in competition. Moreover, the ICC made no attempt to ascertain the extent to which other water carriers will be foreclosed from this traffic; it only assumed the amount would be “small”. Finally, the assumption that other barge carriers will have a nearly equal opportunity to compete with ACL for CSX-generated traffic is belied by the ICC’s heavy emphasis on the supposed “efficiencies” that will result from this merger. If those “efficiencies” are as great as the ICC suggests, other water carriers will not be able to compete effectively with ACL even if given the opportunity; and, it can be predicted, the ICC will rely in the future on this imbalance to justify mergers of other water carriers with railroads, ultimately defeating Congress’s goal of separate ownership of water and rail carriers.
Most significant for the instant case, however, is the potential for predatory pricing designed to drive other water carriers from ACL’s routes. Such an attempt to monopolize would be made possible by the railroad’s ability to “subsidize” reduced rates on water routes with higher rates on all-rail routes. While it is true that an increase in price, made possible by ACL-CSX achieving market power, might encourage other barge operators to enter the market, the real danger lies in driving competitors from ACL’s water routes with price reductions.
The ICC discounted this admittedly real possibility by stating that:
“The barge industry is comprised of a large number of firms, many owned or controlled by large and well-financed corporate parents.... It would not be in the best interests of these large corporate parents to sit back and allow CSX-ACBL to destroy their investments in the barge industry. Instead, they could be expected to exert pressure on CSX, both through antitrust actions and competitive responses, to forego any attempt to monopolize the market_” (Dec. at 36).
Such abdication of responsibility on the part of the ICC makes a mockery of its regulatory functions. It is difficult to fathom this reasoning — by any agency created to regulate an industry of the very highest national priority — which suggests that private enterprise must fill the role entrusted to the ICC, and that private antitrust actions are the most efficient way to preserve the competitive market that Congress intended in enacting the Panama Canal Act. Congress already has recognized the inadvisability of this proposition. Under 49 U.S.C. § 11341(a) (1982), ICC approval of a carrier consolidation immunizes the trans*1200action from the applicability of the antitrust laws. By building competitive principles into the Panama Canal Act, Congress clearly intended not only that the ICC regulate rail/barge consolidations in the public interest but also that the ICC 'police competition in the two industries.
As the ICC itself held in Lake Line Applications Under Panama Canal Act, 33 I.C.C. 699 (1915), in which divestiture of bargelines by the railroads was first ordered:
“These boat lines under the control of the petitioning railroads have been first a sword and then a shield. When these roads succeeded in gaining control of the boat lines which had been in competition with paralleling rails in which they were interested, and later effected their combination ..., by which they were able to and did drive all independent boats from the through lake-and-rail transportation, they thereby destroyed the possibility of competition with their railroads other than such competition as they were of a mind to permit. Having disposed of real competition via the lakes, these boats are now held as a shield against possible competition of new independents. Since it appears from the records that the railroads are able to operate their boat lines at a loss where there is now no competition from independent lines, it is manifest that they could and would operate at a further loss in a rate war against independents. The large financial resources of the owning railroads make it impossible for an independent to engage in a rate war with a boat line so financed.” Id. at 716.
The ICC’s argument that rail/barge competition is irrelevant to the determination of whether a merger will reduce competition on water routes is confounded by the facts and by its own prior decisions.
The majority’s two paragraph discussion of this issue strikes me as blinking at a most serious deficiency in the ICC decision. See majority op. at 1186. While it is true that “[bjarges, not trains, operate on water routes”, in the Act Congress obviously concluded that railroad entry into the barge market might have an adverse effect on competition for water routes. Moreover, in view of the structure of the industries, it was well nigh impossible for the ICC “reasonably” to„conclude that it could disregard rail/barge competition when considering the effects of a merger on a given water route.
It appears that, despite more than 70 years of separate ownership of barge and rail lines, during which time, as the ICC itself stated, “the barge industry has developed into an efficient, effective competitor for intercity freight traffic” (Dec. at 14), if the instant ICC decision is permitted to stand, history will be allowed to repeat itself and the very evil the Panama Canal Act was intended to prohibit will be permitted to flourish.
III.
To summarize:
The ICC has failed to consider the effect of the acquisition on rail/barge competition — as has been the ICC’s practice in prior cases — and has failed to consider the effect that a reduction in rail and water competition would have on purely water carrier competition.
Until Congress sees fit to amend or repeal the Panama Canal Act — something it has repeatedly refused to do — the ICC and the courts are duty-bound to apply it properly. Since I am convinced that the ICC has committed serious errors of law in its interpretation and application of the Act, I would vacate the decision of the ICC and remand the case for further proceedings in accordance with proper legal standards. From the majority’s refusal to do so, I respectfully dissent.
. Throughout this dissenting opinion, the transaction involved is referred to interchangeably as an "acquisition" or "merger".
. Since I believe that the Act has been improperly interpreted and applied by the ICC, this dissenting opinion is addressed solely to that issue, it being neither necessary nor appropriate in my view to reach any other issues.