General Chemical Corporation v. United States

817 F.2d 844, 260 U.S. App. D.C. 121, 1987 U.S. App. LEXIS 5703

• Court: Court of Appeals for the D.C. Circuit
• Decided: May 1, 1987
• Docket: 85-1347
• Status: Published

Opinion

817 F.2d 844

260 U.S.App.D.C. 121

GENERAL CHEMICAL CORPORATION, et al., Petitioners,
v.
UNITED STATES of America and Interstate Commerce Commission,
Respondents,
Atchison, Topeka and Santa Fe Railway Company, et al., Intervenors.

No. 85-1347.

United States Court of Appeals,
District of Columbia Circuit.

Argued April 22, 1986.
Decided May 1, 1987.

Evelyn G. Kitay, I.C.C., with whom Robert S. Burk, Gen. Counsel, and Ellen D. Hanson, Associate Gen. Counsel, I.C.C., Douglas H. Ginsburg, Asst. Atty. Gen. at the time the brief was filed, and John J. Powers, III and George Edelstein, Dept. of Justice, Washington, D.C., were on the brief, for respondents.

Robert B. Batchelder, Omaha, Neb., with whom John M. Edsall, James V. Dolan, Washington, D.C., Louise A. Rinn, Omaha, Neb., John A. Daily, Philadelphia, Pa., John C. Danielson, Detroit, Mich., John Doeringer, Chicago, Ill., Albert Laisy, Baltimore, Md., William L. Phillips, Chicago, Ill., Michael E. Roper, Dallas, Tex., Alice C. Saylor, Pittsburgh, Pa., William H. Teasley, Atlanta, Ga., Stuart E. Vaughn, and Dennis W. Wilson, Chicago, Ill., were on the joint brief, for intervenors.

John K. Maser, III, Washington, D.C., with whom Michael M. Briley, Toledo, Ohio, John F. Donelan, and Richard D. Fortin, Washington, D.C., were on the brief, for petitioners.

Nicholas J. diMichael, Washington, D.C., entered an appearance for petitioners.

Before SCALIA* and SILBERMAN, Circuit Judges, and WRIGHT, Senior Circuit Judge.

Opinion for the court Per Curiam.

PER CURIAM:

Petitioners, producers and receivers of soda ash, challenge the reasonableness of rates assessed by the intervenor railroads for transportation of soda ash produced in Green River, Wyoming. The Interstate Commerce Commission (ICC) decided that the railroads were not "market dominant" as the Railroad Revitalization and Regulatory Reform Act ("4-R" Act) requires for the Commission to have jurisdiction to consider the reasonableness of rail rates. The Commission accordingly dismissed petitioners' complaint. It is the ICC's conclusion about market dominance that is at issue in this case. The Commission based its conclusion of no market dominance on the existence of effective geographic competition in the relevant soda ash market. We find the Commission's analysis of geographic competition to be internally inconsistent and inadequately explained, and thus we conclude that its ultimate finding of no market dominance was arbitrary and capricious and not supported by substantial evidence on the record considered as a whole. Although the Commission's analyses of product and intra- and intermodal competition were not similarly flawed, the Commission was careful not to rest its holding on these sources of competition. They therefore cannot themselves support the conclusion of no market dominance. Hence, the Commission's decision on the issue of market dominance must be vacated and the case remanded to the agency for reconsideration.

I. FACTUAL AND PROCEDURAL BACKGROUND

Soda ash or sodium carbonate is the ninth most widely used chemical in the United States. It is an essential raw material in several industries, particularly the manufacture of glass. There are three methods of producing soda ash. The first is by mining trona ore. This is the method used by the petitioner producers in this case, and it accounts for 80 percent of total United States production capacity. The world's largest deposit of trona ore is located in Green River, Wyoming. The second method of producing soda ash is by distilling spring or lake brine. This method of production is used by facilities located at Searles Lake, California, and it accounted for approximately 11 percent of domestic production capacity in 1982. The third method of producing soda ash is by synthetic production. Although this was at one time the most popular method of producing soda ash, use of synthetic plants has decreased markedly in recent years because of more stringent environmental regulation and increasing energy costs. Synthetic production accounted for only 6.6 percent of total United States production capacity in 1982.

Green River soda ash, the subject of petitioners' challenge, is moved from Wyoming almost exclusively by rail. There is only one originating railroad--Union Pacific--with an average length of haul of over 900 miles. The intervenor railroads' share of total Green River soda ash shipments exceeds 95 percent. The gist of petitioners' complaint is that they are captives of the railroads. Because they have no realistic alternative for transportation of soda ash, they argue, they are forced to pay whatever the railroads choose to charge. Essentially the Green River producers claim that they are surrendering their profits from garnering 80 percent of the national soda ash market to the railroads, who have gained 95 percent of the market for transportation of Green River soda ash.

The 19 petitioners (4 producers of soda ash and 15 receivers) allege in a complaint filed in March 1981 that the rates assessed by the intervenor railroads on Green River soda ash over 238 specified routings to 157 destinations are unreasonably high in violation of the Interstate Commerce Act.1 They seek both prescription of reasonable rates for the future and reparations for past shipments. Forty-nine intervenors (39 railroads and 10 belt or terminal operators) contest this allegation.

The complaint was adjudicated in a bifurcated procedure. Before it can reach the second stage of assessing the reasonableness of rail rates, the ICC must clear two preliminary jurisdictional hurdles. First it must determine that the challenged rates exceed a specified revenue/variable cost ratio, set on a graduated scale chronologically. See 49 U.S.C. Sec. 10709(d)(2) (1982). Second, it must find that the railroads have "market dominance," defined as "the absence of effective competition from other carriers or modes of transportation, for the traffic or movement to which a rate applies." 49 U.S.C. Sec. 10709(a) (1982). Therefore the Administrative Law Judge (ALJ) assigned to the instant case made two separate decisions: in Phase I he decided that the railroads did have market dominance over those movements for which the rates exceeded the revenue/variable cost threshold, but in Phase II he concluded that they nonetheless were not charging unreasonable rates.

This separation of market dominance determinations from assessments of rate reasonableness reflects Congress' conscious rejection of perfect competition as the governing norm in railroad regulation. Rail regulation under the 4-R Act does not require regulation of rates merely because of market imperfection. See H.R.Conf.Rep. No. 781, 94th Cong., 2d Sess. 148 (1976). Rather, the bifurcated procedure permits the conclusion that effective competition is lacking and therefore the railroads have market dominance, but that the rates assessed by the railroads are nonetheless "reasonable." See H.R.Conf.Rep. No. 768, 94th Cong., 1st Sess. 121 (1975). The isolation of the market dominance inquiry thus is central to Congress' plan of structured deregulation in the rail industry.

In deciding whether the railroads have market dominance, the ALJ applied the guidelines promulgated by the ICC in 1981. See Ex Parte No. 320 (Sub-No. 2), Market Dominance Determinations and Consideration of Product Competition, 365 ICC 118 (1981), aff'd sub nom. Western Coal Traffic League v. United States, 719 F.2d 772 (5th Cir.1983) (en banc ), cert. denied, 466 U.S. 953, 104 S.Ct. 2160, 80 L.Ed.2d 545 (1984). Prior to 1981, market dominance determinations were conducted according to rebuttable presumptions based on market share, rail-related investment, and revenue/variable cost ratios. See 49 C.F.R. Sec. 1109.1; P. DEMPSEY & W. THOMS, LAW AND ECONOMIC REGULATION IN TRANSPORTATION 165 & n. 50 (1986). The 1981 guidelines replaced these "on/off" quantitative presumptions with qualitative guidelines that are, in the ICC's words, "broader and more flexible." Market Dominance Determinations, 365 ICC at 119. These guidelines call for the agency to assess the existence of four types of competition: geographic, product, intramodal, and intermodal. Geographic competition is "a restraint on rail pricing stemming from a shipper's or receiver's ability to get the product to which the rate applies from another source, or ship it to another destination." 365 ICC at 128. Product competition exists "when a receiver or shipper can use a substitute(s) for the product covered by the rail rate." Id. Intramodal competition is "competition between two or more railroads transporting the same commodity between the same origin and destination," whereas intermodal competition is "competition between rail carriers and other modes for the transportation of a particular product between the same origin and destination." Id. at 132-33. The railroads bear the burden of identifying where product or geographic competition exists. Once the railroad has made this identification, the shipper in turn has the burden of proving that the product or geographic competition identified by the railroad is not effective.2

In his Phase I decision the ALJ first considered the revenue/variable cost jurisdictional threshold and found that all but eight destinations had revenue/variable cost ratios above the then-current statutory threshold of 160 percent. To arrive at this conclusion the ALJ apparently accepted petitioners' evidence regarding cost. See Phase I Decision at 5. The ALJ then moved to consideration of qualitative evidence of market dominance and concluded that the railroads were market dominant. He first determined that the geographic competition alleged by the railroads was not effective and had not affected the railroad rates from the Green River origins. Id. at 8. He similarly found that the record as a whole did not support a finding that there is effective product or inter- or intramodal competition. Id. at 9, 10-11.

Having concluded that the railroads had market dominance, the ALJ went on to consider the reasonableness of the rail rates. He held that the rates had not been shown to be unreasonable, primarily because the railroads had not yet achieved "revenue adequacy"--i.e., a system-wide return on investment equal to the railroad industry's current cost of capital. Phase II Decision at 14. The ALJ thus refused to prescribe rates for the future and denied reparations. The complaints were dismissed. To reach his conclusion the ALJ apparently accepted the railroads' cost evidence as the best evidence of record. See id. at 15. The railroads appealed to the Commission the Phase I finding that they have market dominance, and petitioners appealed the Phase II finding that none of the assailed rates are unreasonable.

The Commission reversed the ALJ's finding of market dominance. As a threshold matter, it adopted the railroads' cost evidence for determination of the statutory revenue/variable cost ratio and thus dismissed several more complaint movements with regard to reparations and/or future rate prescription. See ICC Decision at 5. It then held that it lacked jurisdiction to consider the reasonableness of the remaining rates because "the totality of the evidence shows that there is competition for [soda ash] traffic moving between the many points involved in the complaint and that this competition effectively constrains the railroads from charging unreasonable rates on the traffic at issue." Id. at 7-8 (footnote omitted). The Commission based its holding primarily on its finding of "substantial" geographic competition from three alternative sources of soda ash, which in its eyes was "strong enough, standing alone, to warrant a finding that there is no market dominance." Id. at 9. It also found that product competition "enhances" and "fortif[ies]" its finding of effective geographic competition, and that evidence of intra- and intermodal competition "strengthens [its] view that market dominance is lacking." Id. at 14-15. The Commission thus concluded that it did not have jurisdiction over the rate reasonableness issues reached by the ALJ below.3 The petitioners have appealed this final decision by the agency.

II. THE STANDARD OF REVIEW

Our review of the ICC's decision that the intervenor railroads are not market-dominant is governed by Section 10 of the Administrative Procedure Act, 5 U.S.C. Sec. 706 (1982). The APA requires us to vacate the agency's decision if it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law * * * [or] unsupported by substantial evidence." Id. Sec. 706(A) & (E). The Supreme Court has explained that "[t]o make this finding the court must consider whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment. Although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not empowered to substitute its judgment for that of the agency." Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 417, 91 S.Ct. 814, 824, 28 L.Ed.2d 136 (1971) (citations omitted). This standard of review ensures that the agency has engaged in reasoned decisionmaking, see Cross-Sound Ferry Services, Inc. v. ICC, 738 F.2d 481, 483 (D.C.Cir.1984), that is both adequately explained, see Center for Science in the Public Interest v. Dep't of the Treasury, 797 F.2d 995, 999 (D.C.Cir.1986), and supported by substantial evidence in the record as a whole. See Universal Camera Corp. v. NLRB, 340 U.S. 474, 488, 71 S.Ct. 456, 464, 95 L.Ed. 456 (1951).

III. COST EVIDENCE AND THE STATUTORY REVENUE/VARIABLE COST RATIO

Petitioners challenge as arbitrary and capricious the Commission's rejection of their cost evidence for assessing revenue/variable cost ratios. As noted above, the Commission must find that the railroads' operations on a given route exceed a statutory revenue/variable cost threshold before it has jurisdiction to reach the issue of market dominance. See 49 U.S.C. Sec. 10709(d)(2) (1982). The Commission concluded that the railroads had presented the best evidence of cost. It thus dismissed several complaint movements for failure to meet the then-current statutory threshold of 1.6. Although the Commission rejected petitioners' cost evidence across the board, it explicitly objected to only two specific areas within those calculations. Petitioners challenge the Commission's reasoning in both areas.

A. The Rail Car Cost Controversy

Rail car costs come in two varieties: on-line costs, incurred by a railroad when it uses its own cars to transport cargo on its own lines, and off-line costs, incurred when a road leases cars from another road or from private investors. At issue is the proper method of calculating off-line car costs. The Commission rejected petitioners' "annuity" method for arriving at these costs on the ground of inadequate (and illegible) documentation. Petitioners respond that the Commission's action was arbitrary and capricious because it was based on review of only part of the documentation present in the record.

In lieu of petitioners' annuity-based evidence, the Commission accepted the railroads' evidence of off-line costs, which are based on the per diem charges paid by lessee roads to lessor roads for the use of rail cars. Petitioners object to this evidence because they feel it overstates the actual cost to the lessor roads. They point out that the per diem charge includes some margin of profit for the lessor, and that the charge is therefore inaccurate as a measure of costs. Be that as it may, the per diem method is the traditional approach to assessing the costs of off-line rolling stock. Absent a stronger showing of unreasonableness, the court will not label that method per se "arbitrary and capricious."

In addition, petitioners argue that the Commission failed to consider all the evidence on the record. Petitioners claim that the Commission ignored documents that clearly established the process by which the annuity-based cost figures had been calculated. The record does not bear out this charge. In fact, a close examination of the ICC's Financial Analysis Section memorandum on the matter reveals several comments on precisely the documents petitioners claim the Commission ignored. See ICC Memorandum, June 18, 1984, at 4, Joint Appendix (JA) 1312, referring to Verified Statement of William W. Whitehurst, Jr., Workpapers 566, 569, JA 1198-99. Clearly, those documents were considered before any final decision was made on the overall usefulness of petitioners' cost calculation evidence.

Finally, petitioners argue that the Commission acted unreasonably in rejecting petitioners' cost evidence because of illegibility without giving them the opportunity to submit clarifying material, and in requiring them to offer an additional "narrative" explanation of their annuity calculations. Although petitioners may find the Commission's requirements burdensome, the ICC clearly has the power to determine what constitutes acceptable and unacceptable legibility and elaboration in evidence presented, and to penalize those who submit documents that fall on the wrong side of that line, as long as it does not do so in a discriminatory fashion. See Airmark Corp. v. FAA, 758 F.2d 685, 691-95 (D.C.Cir.1985). It is difficult to see how the Commission's actions on these matters could be deemed discriminatory in this case, and we thus refuse to find them arbitrary and capricious.

B. The Maintenance of Way Costs Controversy

The controversy in this area centers around the use of two formulas for maintenance costs: the venerable "Rail Form A" formula (RFA), and the new "Speed Factored Gross Tonnage" formula (SFGT). The ALJ below accepted petitioners' use of the SFGT formula in the Phase I determination of market dominance. See Phase I Decision at 8. But when time came to consider the reasonableness of the rail rates in Phase II, the ALJ rejected the SFGT formula on the ground that "some of the co-efficients expressed in the formula are not verifiable." Phase II Decision at Appendix C, p. 11. The Commission adopted the ALJ's Phase II reasoning and thus rejected the SFGT approach not only for any Phase II considerations, but also for the Phase I market dominance determination. ICC Decision at Appendix D, p. 3.

Petitioners have offered no evidence that the Commission failed to act reasonably in choosing the controlling formula. The merits and demerits of the two competing formulas are matters for the Commission's expertise, not ours. We decline petitioners' invitation to attempt an independent assessment of this matter. Because reasonable Commissioners might differ on the appropriate formula, we defer to the ICC's expertise on this second aspect of the cost controversy as well.

IV. MARKET DOMINANCE

As noted above, the ICC in 1981 replaced its quantitative rebuttable presumptions of market dominance with a qualitative test centering on presentation of evidence of four types of competition. The Commission offered general guidelines for submission of evidence regarding the effectiveness of these types of competition. Unfortunately, it has failed in this case to explain adequately its conclusions on competition and market dominance. The Commission based its holding of no market dominance primarily on its finding of effective geographic competition, but it failed to adhere to its own guidelines regarding evidence of geographic competition and failed to explain satisfactorily its conclusion that geographic competition was effective in light of the record as a whole. The Commission's more cursory analyses of product, intra- and intermodal competition, while not similarly flawed, were not held by the Commission to be sufficient to sustain its ultimate conclusion in the absence of a finding of effective geographic competition. The agency's decision thus falls below the standard of reasoned decisionmaking. We cannot say, and it is not our role to say, that the railroads are market dominant. We do not hesitate to conclude, however, that the ICC's decision that the railroads are not market dominant is arbitrary and capricious and not supported by substantial evidence on the record considered as a whole. The ICC's decision is therefore vacated and the case remanded to the agency for further consideration and much-needed elucidation.

A. Geographic Competition

In its 1981 Market Dominance Determinations the ICC offered the following guidelines for submission of evidence concerning geographic competition:To establish the potential for geographic competition, evidence should be submitted concerning the following: (1) the number of alternative geographical sources of supply or alternative destinations available to the shipper or receiver for the product in question; (2) the number of these alternative sources or destinations served by different carriers; and (3) that the product available from each source or required by each destination is the same.

Such evidence is sufficient only to indicate whether effective geographic competition is possible. To determine whether effective geographic competition actually exists, evidence showing the feasibility of each source or destination and the likelihood of competition should be presented.

365 ICC at 134 (emphasis added). In its opinion, however, the Commission failed to account for evidence in the record that casts doubt on the existence of the second and third evidentiary criteria for a prima facie showing of geographic competition. In fact, the Commission ignores its own guidelines, of which it makes no mention, and the record evidence that undermines its finding of market dominance.

The Commission has no difficulty finding that the railroads cleared the first evidentiary hurdle. It notes the uncontroverted existence of three alternative sources of soda ash: Searles Lake, California, which refines lake brine to make soda ash, and Solvay, New York and Amherstburg, Ontario, Canada, which manufacture soda ash synthetically. The Commission's treatment, however, of the second and third criteria in its guidelines for a showing of geographic competition is seriously flawed. The second criterion is a showing of "the number of * * * alternative sources * * * served by different carriers." 365 ICC at 134. The purpose of this criterion is apparent: if the same carrier serves both the complainant source and the alternative sources, the alternative sources might not provide geographic competition because the railroad might be indifferent to competitive "losses." Normally, when geographic competition exists the receivers of soda ash can simply order their soda ash from a different supplier if rail rates from the complainant source are too high. This possibility of losing shipping business will, in theory, keep the railroad that serves the complainant source from charging excessive rates to begin with. But if the same railroad serves the alternative sources as well as the complainant source, the railroad presumably will be indifferent to the loss of business from shipping the complainant source's product, because the business is not really lost, just shifted to another of the railroad's clients. The competitive restraint on the railroad's rates is thus undermined.

The Commission seems to refer to this aspect of its guidelines in its opinion by noting that "[t]he soda ash from these alternative origins moves by different transportation than the soda ash from Green River." ICC Decision at 10. The Commission goes on to cite the railroads' own evidence to list the various rail carriers serving the alternative origins. However, it neglects to note that many of these carriers also serve the Green River complainants. The only transportation listed by the Commission for Solvay soda ash is Conrail. But Conrail is one of the defendants named by the Green River producers. In fact, Conrail's average length of haul of Green River soda ash is 850-900 miles--almost as long as that of Union Pacific, the sole originating railroad for Green River soda ash. See Brief of Petitioners at 10. The Commission notes that the alternative source at Amherstburg is served by "the Essex Terminal Railway, a short line that connects at Windsor, Ontario with five major railroads." ICC Decision at 10. The Commission fails to mention that three of these five railroads are defendant railroads in this case (Conrail, C & O, and N & W). The Commission notes that Searles Lake soda ash moves by ocean vessel and motor carriers as well as by rail. But its rail carrier is the Trona Railway Company, a 31-mile railroad owned by Kerr-McGee, which brings it to an interchange with the Southern Pacific--another defendant railroad in this case.

The Commission's assertion that the alternative origins are served by different carriers is suggestive of the considerations outlined in its guidelines but does not adequately explain its decision in light of them. The Commission's inexplicable failure to make any mention of the overlap between the carriers it lists to "support" its assertion and the carriers of Green River soda ash falls short of the reasoned decisionmaking required of the agency. It may well be that some of the rail carriers that serve the alternative sources are independent of the Green River carriers, or that the different carriers still compete with each other even though they also serve alternative sources, or that the alternative sources are served by means other than rail. But the Commission advances none of these explanations. It simply lists without explanation the rail carriers serving the alternative sources, seemingly oblivious to their demonstrable connection to Green River. The Commission's own guidelines and judicial precedent clearly establish that any overlap between the carriers to a complainant source and those that serve allegedly competitive sources is relevant to a determination of the existence of geographic competition. See Arizona Public Serv. Co. v. United States, 742 F.2d 644, 654 (D.C.Cir.1984). The Commission impermissibly failed to consider this relevant factor in its opinion. See Motor Vehicle Mfrs. Ass'n v. State Farm Mutual Auto Ins. Co., 463 U.S. 29, 46-57, 103 S.Ct. 2856, 2868-74, 77 L.Ed.2d 443 (1983).

The third evidentiary criterion in the guidelines on geographic competition is a showing "that the product available from each source * * * is the same." The Commission seems to refer to this aspect of its guidelines by concluding in a footnote that "[t]here is no evidence * * * that synthetic soda ash [produced at Solvay and Amherstburg] is not essentially the same product as natural soda ash. Accordingly, we will treat synthetic soda ash as a close substitute for what is being mined and processed at Green River." ICC Decision at 9 n. 13. But there is much evidence on the record that Green River produces primarily dense soda ash, whereas the synthetic plants at Solvay and Amherstburg produce both light and dense ash. See Verified Statement of T.J. Kessler at 21, JA 108; Verified Statement of David A. Westerlund at 7, JA 220; Rebuttal Verified Statement of Robert H. Richards at 23, JA 766; Rebuttal Verified Statement of Donald E. Watson at 4, JA 780. Moreover, one of the railroads' own witnesses conceded during cross-examination that light and dense soda ash differ in terms of their use:

Q: You are not comparing dense and light? You are not equating the two, are you, sir?

A: No. But if something was locked in, if some industry was locked into the use of light ash in the past, it seems as though an adjustment either in their plant or in a product in Wyoming could be made to use dense ash rather than to be locked into the use of light ash.

Q: But the, to use your word, "adjustment" would have to be made somewhere along the line, either in the type of ash that is produced or refined, I suppose the term is, or, as a raw material, what it is being used for at a given destination?

A: Sure.

Oral Testimony of Railroad Witness Foley, JA 886-87. We cannot determine whether the Commission considered this evidence or whether it impermissibly ignored it in concluding that there is "no evidence" that synthetic soda ash, some of which is light ash, is not "essentially the same product" as Green River mined soda ash. The Commission failed to explain what it meant by "essentially the same product" or "close substitutes." To the extent that light and dense ash differ, the light soda ash produced at Solvay and Amherstburg is not directly in competition with the dense soda ash produced at Green River, and one of the three criteria for a finding of geographic competition is not met. The Commission's ultimate conclusion--that natural and synthetic soda ash are close substitutes--may be correct. But the Commission's failure lies in neglecting to address directly the testimony on the record about the differences between light and dense soda ash in order to explain its conclusion that the two are "essentially" the same. See Amoco Oil Co. v. EPA, 501 F.2d 722, 741 (D.C.Cir.1974) ("Where [the agency's actions] turn crucially on factual issues, we will demand sufficient attention to these in the statement to allow the fundamental rationality of the [actions] to be ascertained.").

In addition to neglecting its own evidentiary guidelines to determine whether geographic competition is even possible, the Commission also failed to explain adequately its conclusion that geographic competition is effective. According to the Commission's guidelines, the railroads bear the burden in a market dominance proceeding of identifying sources of geographic and product competition. The shippers then bear the burden of demonstrating that the identified competition is not effective. 365 ICC at 132. But the railroads offered no evidence of geographic competition at 27 of the 157 destinations complained of by the petitioners. See Defendants' Initial Submission Identifying Sources of Geographic and Product Competition, JA 54; Brief of Petitioners at 30 n. 1 (listing by number the destinations for which defendants identified no geographic competition). Five of these 27 destinations were later dismissed by the Commission for falling below the statutory revenue/variable cost ratio. See ICC Decision at Appendix C, pp. 1-9 (destinations 30, 41, 112, 140, and 142 were dismissed). That leaves 22 destinations or 14 percent of the total number unaccounted for. The Commission makes no mention of these destinations and fails to explain how its conclusion that geographic competition is "strong enough standing alone" to support its holding can be maintained in light of them.

We do not suggest that the Commission must consider each complainant destination individually. It is within the Commission's discretion to use a "grouping approach," as it explained in its opinion, "differentiat[ing] among different routes or movements only to the extent that the factual assertions and legal and economic arguments raised by the parties are not common to them all." Id. at 8 n. 12. The problem is that the Commission did not adhere to its own approach. Clearly, the 22 destinations for which no evidence of geographic competition was offered by the railroads should be treated separately from the other destinations according to the Commission's own criteria. These unaccounted-for destinations deserve at least some mention by the Commission, since it relies almost exclusively on its finding of geographic competition.4

Moreover, the Commission ignored the defendants' own acknowledgment that geographic competition is not effective for 72 destinations or 46 percent of the total number. The ALJ asked the parties to submit briefs for his Phase I decision in the form of proposed opinions. Defendants' brief conceded the lack of effective geographic competition at almost half of the complainant destinations.5 The Commission's opinion did not mention this concession by the defendants. This omission is disturbing because the opinion rests upon a conclusion--that geographic competition alone could support the Commission's finding--that even the railroads did not urge. When the agency wishes to base its conclusions on arguments not endorsed even by the party in whose favor it rules, it owes a duty of adequate explanation. The Commission's failure even to note its divergence from the railroads' own contentions prevents us from being able reasonably to discern "the agency's path." Bowman Transportation v. Arkansas-Best Freight Sys., 419 U.S. 281, 286, 95 S.Ct. 438, 442, 42 L.Ed.2d 447 (1974).

In an attempt to explain their apparent concession, the railroads argue in their brief on appeal that Appendix A to their proposed opinion contained evidence of rate concessions, labelled there as intramodal competition, which should be considered evidence of geographic competition as well. They note that the Commission treated it as such. Joint Brief of Intervenor Railroads at 30 n. 21. There are two problems with this way of rationalizing the railroads' concession on the geographic competition point. The first is that the railroads consistently argued the rate concession point as intramodal evidence, see Joint Verified Statement of Richard J. Barker and Thomas E. Schick at 10-20, JA 263-74; Verified Statement of John R. Sunnygard at 3-16, JA 501-14; Defendants' Proposed Initial Decision on Market Dominance at 52, JA 931, and in any case the text of their submission should overcome any contrary suggestion in an appendix.

The second problem runs deeper and points to an internal inconsistency in the Commission's opinion. It is true that the Commission did cite evidence of rate concessions and mere cost recovery rate increases (as opposed to revenue-based increases) as evidence of geographic competition. See ICC Decision at 10 & nn. 15 & 16, 11. But the Commission earlier in its opinion denied petitioners' request for official notice of rate increases to certain destinations because such evidence dealt only with intramodal competition (which the Commission found was not effective) and therefore was merely cumulative. Id. at 4. The Commission cannot have it both ways. If evidence regarding rate concessions or increases is probative of geographic competition, as the Commission maintains at one point, then petitioners' request for official notice should have been granted. If such evidence is solely relevant to intramodal competition, as the Commission maintains elsewhere in its opinion, then the Commission should not have cited it to support its conclusion regarding geographic competition. The Commission is free to regard evidence of rate concessions as evidence of either geographic or intramodal competition or both. But it must do so in a rational and consistent manner that is fair to the parties involved.

Finally, the Commission supports its finding of effective geographic competition by concluding without explanation that "Searles Lake is the low cost producer of soda ash." Id. at 9. The low cost producer controversy is quite significant to the issue of geographic competition. Essentially, petitioners claim that their cost advantage over other producers of soda ash is being appropriated by the railroads. The railroads and the Commission respond to this claim in part by citing evidence that the delivered price of soda ash from alternative sources is competitive with that from Green River. See Joint Brief of Intervenor Railroads at 30-36; ICC Decision at 10. But this focus on delivered prices is unresponsive to petitioners' claim. Petitioners argue that the railroads are appropriating through excessive rates the profits that the petitioner producers would ordinarily reap from their low cost soda ash. Because the delivered prices already include the amount paid to the railroads, the fact that other sources can sometimes offer competitive delivered prices does not determine whether the rates the railroads charge are excessive. It could be that the railroads are simply effective monopolists, already charging the maximum rates that their monopoly will bear. Evidence that delivered prices from alternative sources are competitive can be evidence of geographic competition only if Green River is not the low cost producer. Thus yet another aspect of the Commission's opinion--its emphasis on competitive delivered prices--depends on its unexplained assertion that Searles Lake is the low cost producer of soda ash.

Only one witness testified that Searles Lake is in fact the low cost producer, and even his testimony was qualified. See Verified Statement of John R. Sunnygard at 31, JA 529 (Union Pacific's marketing manager testified that "Kerr-McGee's marginal cost of production is believed to be the lowest in the industry") (emphasis added). Moreover, his testimony that Searles Lake ash was cheaper to produce than Green River ash because it could be pumped out of the lake rather than mined was rebutted by testimony from petitioners' witnesses that suggested that the cost of refining Searles Lake ash after it was pumped up raised its price to above that of Green River ash. See Verified Statement of Donald R. Krause at 23, JA 137; Verified Statement of Robert H. Richards at 6, JA 181. In addition, petitioners note that it is uncontradicted that FOB origin prices from Green River are lower than those from Searles Lake. See Brief of Petitioners at 45 ($107.25 NT effective Feb. 1983 FOB Searles Lake as compared to $84 NT effective Feb. 1983 FOB Green River). Furthermore, the interrogatory answers from petitioners available under seal, which the railroads cite to show that delivered prices are sometimes competitive, also show that FOB prices from Green River are uniformly the lowest. The Commission simply cites the railroads' one witness and ignores the uncontradicted FOB price evidence offered by petitioners. And later in its opinion the Commission seems to softpedal its low cost producer point by noting that "the fact that * * * that product [Green River soda ash] represents a slight cost advantage to a particular shipper * * * does not mean that the use of soda ash from other sources is precluded." ICC Decision at 12 (emphasis added). The Commission's failure to explain why the FOB prices are not probative in light of the very meager evidence on the records that supports its low cost producer statement leads us to conclude that this statement is not supported by substantial evidence.

Any one of these inconsistencies or failures of explanation in the Commission's analysis of geographic competition might not require remand. But the collective impact of these problems in light of the fact that the Commission relied heavily if not exclusively on its finding of geographic competition leaves us no alternative to vacating and remanding the Commission's decision. The confusing and inconsistent analysis it offered was so incomplete and conclusory as to fall below the standard of reasoned decisionmaking. As Judge Wilkey has pointed out: "Such intuitional forms of decisionmaking, completely opaque to judicial review, fall somewhere on the distant side of arbitrary." Central Florida Enterprises, Inc. v. FCC, 598 F.2d 37, 50 (D.C.Cir.1978).

B. Product Competition

The Commission rested its conclusion of no market dominance on the existence of geographic competition. ICC Decision at 12. The ICC also found, however, that soda ash faces product competition from a number of sources. While the Commission never indicates that these sources by themselves actually provide "effective" competition sufficient to support a finding of no market dominance, its opinion does state that product competition "fortifies" the finding of effective geographic competition. Standing alone, then, the Commission's product competition findings cannot support the conclusion that no market dominance exists here. In light of the limited nature of the Commission's determination in this area, and in light of its generally adequate explanation of its product competition decisions, we affirm its findings as to product competition, but nonetheless vacate its decision as a whole.

1. Competition from cullet and caustic soda.

The Commission overturned the ALJ's finding that cullet (waste glass) and caustic soda (a chemical that can substitute for soda ash in the manufacture of glass) do not provide significant competition for soda ash. ICC Decision at 13. Petitioners respond that because neither of these materials can replace more than 20-25 percent of the total soda ash demand, they cannot have any significant disciplinary effect on soda ash prices. Brief of Petitioners at 51-52. As a matter of economics, of course, this is incorrect, as the Commission points out. Brief of Respondents at 39.

The record is replete with evidence that caustic soda and cullet are used extensively, and that they are increasingly popular soda ash substitutes. See, e.g., Joint Verified Statement of Richard J. Barber and Thomas E. Schick at 50-64, JA 312-27. Importantly, the ICC does not say that cullet and caustic soda provide effective competition in themselves. It has merely indicated that these materials do compete to some significant extent with soda ash. Given this cautious conclusion, there is no justification for overturning the finding of limited product competition from these two materials.

2. End-product competition.

Some 33 percent of soda ash consumed in the United States each year goes toward the manufacture of glass containers. The Commission found that competition between glass containers and containers made of paper, aluminum, and steel has a "significant" disciplinary effect on the price of soda ash. ICC Decision at 12-13. In so doing, the Commission reversed the ALJ's decision that end-product competition was not an appropriate consideration in market dominance determinations. Phase I Decision at 9.

Petitioners counter with two arguments. First, they point out that the Commission's own guidelines on market dominance determinations state that "product competition occurs when a receiver or shipper can use a substitute(s) for the product covered by the rail rate." 365 ICC at 128. Aluminum cans and paper bottles are obviously not direct substitutes for the product covered by the rail rate at issue here--soda ash. Second, they argue that because only 55 percent of the complainant destinations manufacture non-glass containers, such containers cannot offer "effective" competition. Brief of Petitioners at 49. Neither argument is persuasive.

The ICC's guidelines do not clearly allow for consideration of end-product competition. Nevertheless, if we read the Commission's decision in the present case as finding that soda ash competes with aluminum and paper pulp, rather than with aluminum cans and paper bottles (a reasonable reading), the decision falls generally within the scope of this circuit's decision in Aluminum Co. of America v. ICC, 761 F.2d 746 (D.C.Cir.1985). In that case the court found that the Commission could consider the competition facing aluminum cans from plastic and steel containers in determining the existence or nonexistence of market dominance in aluminum rail rates. 761 F.2d at 751. Granted, Aluminum Co. is distinguishable from the present case in that soda ash is merely a raw material for production of glass, not glass itself. In Aluminum Co. the material at issue was aluminum itself. A case truly analogous to the present case would involve shipment of some component of aluminum, bauxite for example, rather than aluminum.

Although this aspect of the present case distinguishes it from Aluminum Co., it is unclear that the distinction makes different legal conclusions necessary in the two actions. Unlike glass, aluminum is not manufactured at the same site where it is transformed into a final product. As a consequence, bauxite and other components of aluminum simply are not shipped to aluminum can manufacturing plants. Aluminum ingots are. If the price of soda ash drives up the price of glass to noncompetitive levels, container manufacturers' demand for soda ash will drop and their demand for aluminum ingots--not bauxite--will rise. Bauxite does not compete directly with soda ash for transport to container plants, aluminum does. Viewed in this light, Aluminum Co. seems squarely on point with the present case. The relevant substitute for soda ash, the "product covered by the rail rate," is aluminum ingot, not some component of aluminum.

If we assume that product competition is a valid consideration in the determination of market dominance, as even petitioners seem to assume, end-product competition seems an appropriate consideration as well. It may well be that some kinds of end-product competition are simply too indirect and attenuated influences to be an appropriate part of the market dominance determination. Here, however, petitioners only half-heartedly challenge the use of end-product competition as a factor in the overall calculation. They argue more forcefully that the record does not support a finding of significant competition from end products. Brief of Petitioners at 49-50. But as the Commission has not heavily relied on the end-product competition point, and since evidence on the record does tend to show significant container market competition, petitioners' arguments are not powerful enough to support reversal or remand of the Commission's finding on this point.

C. Inter- and Intramodal Competition

The Commission agreed with petitioners that intra modal competition, i.e., competition among railroads for the traffic in question, is virtually nonexistent. ICC Decision at 14. The Commission found, however, that inter modal competition, for example motor carriage and rail-motor combination arrangements, did have some disciplinary effect on rail prices. Id. at 15. The Commission even hedged a bit on this cautious finding by stressing that intermodal competition "is insufficient to support a finding of effective competition in this case." Id.

Petitioners protest that the record does not support a finding of any competition from motor carriage transport. They would have a colorable argument if the Commission had tried to make much of the intermodal competition finding. Evidence of competition from truck and railtruck transport is very weak, given the quantity of soda ash that must be transported, the enormous conversion costs from motor carriage to rail and back again, and the general absence of any viable backhaul (return cargo). Id. at 16. But the Commission does not base its market dominance determination on this sort of competition. As a consequence, reversal of this exercise of the Commission's judgment is inappropriate. The Commission has adequately supported its inter- and intramodal competition determinations.

In sum, nothing in the Commission's findings on product competition or inter- and intramodal competition warrants reversal. This conclusion rests in substantial part, however, on the fact that the agency very carefully refused to find that these sources themselves provided "effective competition." Clearly, if the Commission's analysis of geographic competition is inadequate, these additional sources of competition will not support its decision independently, in light of its cautious holding. If the Commission were to assert that they do, we would, of course, undertake a much more careful examination of the evidentiary basis for the determinations to ensure that its conclusion is actually supported by substantial evidence.

V. CONCLUSION

Because the ICC's analysis of geographic competition is internally inconsistent and inadequately explained, we find its ultimate conclusion of no market dominance to be arbitrary and capricious and not supported by substantial evidence on the record considered as a whole. Accordingly, the Commission's decision is vacated in part and remanded for reconsideration consistent with this opinion.

It is so ordered.

* Circuit Judge SCALIA was a member of the panel at the time this case was argued, but did not participate in the decision. He became an Associate Justice of the Supreme Court of the United States on September 26, 1986

1 After the petition for review was filed, this court granted the motion of four complainants to withdraw from the case. There are 15 complainants remaining

2 In a decision subsequent to its final decision in this case, the ICC transferred the full burden of proof of product and geographic competition to the railroad. See Ex Parte No. 320 (Sub-No. 3), Product and Geographic Competition (decided Oct. 24, 1985). This decision does not affect the case before us, because there is no reason to believe that the change in policy was meant to be retroactive. On the contrary, the Commission was clearly aware of the pendency of the later case during its consideration of the instant case, see ICC Decision No. 38412S at 17-18 (Commissioners Lamboley and Strenio, concurring in part and dissenting in part, and Commissioner Simmons, dissenting), but decided not to stay its decision. We review the Commission's opinion in accordance with the burdens of production and persuasion in effect at the time of its decision

3 The Commission did note that one aspect of the ALJ's rate reasonableness decision was "glaringly in error." ICC Decision at 16. It rejected the ALJ's suggestion that no rate on captive traffic can be found unreasonable if it is charged by a railroad that the Commission has determined to be revenue inadequate. The Commission noted that "revenue adequacy is not the sole standard for determining the reasonableness of rates and has never been interpreted as such by the Commission. Thus, the fact that a carrier is not revenue adequate does not mean that all of its rates are necessarily reasonable." Id

4 At oral argument counsel for the Commission suggested that this point was waived by petitioners because they did not argue market dominance location-by-location rather than nationwide. But our criticism of the Commission's opinion does not depend on what the parties argued to it. The Commission failed to make a reasoned decision. It explained that it would "group" destinations in order to make an overall judgment about market dominance, apparently did so, and concluded that the existence of geographic competition could support its conclusion that there was no market dominance. But in this chain of reasoning the Commission neglected to notice that the record contains no evidence of geographic competition for 22 destinations. No explanation is offered for this particular "group" of destinations. We do not suggest that there is no possible way to support the agency's ultimate conclusion, but simply that reasoned decisionmaking requires some explanation to fill this rather obvious gap. This requirement of rationality cannot be waived by the parties. As the Supreme Court has stated: "It will not do for a court to be compelled to guess at the theory underlying the agency's action." SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995 (1947)

5 See Defendants' Proposed Initial Decision on Market Dominance at 32, JA 922:

I disagree with defendants that effective geographic competition exists for all of the destinations they have designated. After a careful review of the evidence, I find that effective geographic competition exists for the complaint destinations located in the areas discussed below, and accordingly dismiss the complaint as to those routes and destinations.

The brief/proposed opinion goes on to find effective geographic competition at only 85 destinations.