Opinion for the Court filed by Circuit Judge, STEPHEN F. WILLIAMS.
STEPHEN F. WILLIAMS, Circuit Judge:
Montana wheat and barley farmers ship their grain via Burlington Northern to ports in the Pacific Northwest for transportation to markets in the Pacific Rim. The Interstate Commerce Commission, in a decision not contested here, has found Burlington to have “market dominance” over these shipments; as a result, the Commission has jurisdiction over the rates. See 49 U.S.C. §§ 10701a(b)(l) and 10709(b) (1988). In the set of decisions under review the Commission awarded the shippers over $16 million for excess charges. Both the railroad and the shippers (including the Montana Department of Agriculture as one of their representatives) petition for review. The principal issue revolves around the Commission’s decision not to employ its “constrained market pricing” methodology (“CMP”), which implements the principle of Ramsey pricing and which the ICC has apotheosized as the “preferred and most accurate procedure available for determining the reasonableness” of rates in markets where a carrier has market dominance. See, e.g., McCarty Farms, Inc. v. Burlington Northern, Inc., 3 I.C.C.2d 822, 840 (1987). Instead, the Commission used the “revenue over variable cost” (“R/VC”) method, which holds the railroad’s charges on the disputed traffic to the same ratio of revenue to variable cost as on certain “benchmark” traffic that the Commission has deemed comparable. Burlington claims that the Commission failed to justify this substitution, and we agree. Before we reach that issue, however, we must address claims as to the scope of our jurisdiction and the scope of the complaints filed with the Commission. After dealing with these and with the fight over CMP and R/VC, we address attacks advanced by the shippers that concern the way the Commission applied the R/VC methodology in this case. We address these R/VC issues so that, if the Commission on remand should lawfully adhere to that methodology, the entire litigation may be brought to an end.
Under review are three ICC decisions challenged by both sides and one chai[4]*4lenged by McCarty Farms only, all four decisions being styled “McCarty Farms, Inc. v. Burlington Northern, Inc.”. For simplicity of citation we label them McCarty Farms I through McCarty Farms IV: (1) McCarty Farms I, 4 I.C.C.2d 262 (1988); (2) McCarty Farms II, Docket Nos. 37809, 37809 (Sub-No. 1) and 37815S (unpublished decision served Feb. 21, 1989); (3) McCarty Farms III, Docket Nos. 37809, 37809 (Sub-No. 1) and 37815S (unpublished decision served Mar. 27, 1991); (4) McCarty Farms IV, Docket Nos. 37809, 37809 (Sub-No. 1) and 37815S (unpublished decision served Nov. 26, 1991).
These decisions found Burlington’s rates excessive only on trainload (52-car) shipments of export wheat and barley sent from Montana origins to ports on the Pacific Northwest Coast, and awarded reparations for the years 1981-86, plus interest to July 1, 1991. See McCarty Farms IV, slip op. at 3. The Commission found no overcharges on either single-car or multiple-car (26-car) shipments of wheat or barley. See McCarty Farms III, slip op. at 8.
I.Jurisdiction
McCarty Farms (and the shippers it represents) challenge this court’s jurisdiction to review many of the claims raised in this suit. To assess the argument, we must divide the total dispute into three segments — the first over which we clearly do not have jurisdiction, a disputed middle ground, and the last over which we clearly do have jurisdiction.
1. Single-car shipments of wheat for the two-year period ending September 12, 1980. McCarty Farms started this dispute’s crawl through the legal system in 1980 by filing a class action on behalf of Montana farmers under 49 U.S.C. § 11705(c)(1) and 28 U.S.C. § 1337 in the U.S. district court for the District of Montana. The suit disputed only wheat shipments sent in single cars over a two-year period ending September 12, 1980. Under the doctrine of primary jurisdiction, the district court referred the rate reasonableness question to the ICC, but retained jurisdiction over the lawsuit. McCarty then filed a complaint with the Commission encompassing this claim (plus others, as discussed below). See ICC No. 37809 (filed Mar. 27, 1981).
Because 28 U.S.C. § 1336(b) gives a referring district court exclusive jurisdiction over appeals from Commission orders “arising out of such referral,” 1 all agree that any appeal as to the single-car wheat shipments moving before September 12, 1980 lies in the district court for the District of Montana.
2. Single-car shipments of barley, and wheat shipments after September 12, 1980. McCarty’s complaint to the ICC following up the district court referral, ICC No. 37809, challenged not only Burlington’s rates for single-car shipments of wheat in the original two-year period but also those rates for barley. McCarty also broadened its request for relief to include prospective rate prescriptions in addition to its previous demand for reparations. The ICC assigned the barley rate challenges to a separate proceeding, ICC No. 37809 (Sub-No. 1), because it concluded that barley rates were not a part of the district court’s referral order.
3. Rates on multi-car (26-car) and trainload (52-car) shipments of wheat and barley. The Montana Department of Agriculture also filed a complaint, ICC No. 37815S, attacking Burlington’s rates on larger — multi-car or trainload — shipments .of wheat and barley. See id. at 1. McCarty concedes that § 1336(b) has no application to these rates (because they arise out of the Montana Department of Agriculture complaint, not the district court referral), so that this court has jurisdiction to review the ICC decisions under the Hobbs Act, 28 [5]*5U.S.C. §§ 2321(a) and 2342(5).2
Given the agreement over jurisdiction for rates in categories 1 and 3, the dispute thus relates only to rates in category 2. McCarty argues that the district court for the District of Montana has exclusive jurisdiction not only over the ICC disposition of the claims actually referred to the ICC by that court, but also over the claims that McCarty added when it filed its complaint with the Commission. Although no one here addresses the point explicitly, McCarty also implicitly makes the narrower claim that the district court has exclusive jurisdiction to review the ICC decisions as they relate to reparations for single-car wheat shipments moving between September 12, 1980 and the Commission’s ruling on the reasonableness of those rates.
McCarty’s argument has two steps. First, as we have seen, 28 U.S.C. § 1336(b) gives a district court that refers an issue to the ICC exclusive appellate jurisdiction over any ICC order “arising out of such referral.” McCarty argues — or, more precisely, implicitly asks us to assume — that all the determinations generated by its complaint to the Commission can be said to have “aris[en] out of” the referral, even though that complaint was broader than the district court referral from McCarty’s original complaint. Second, McCarty argues that § 214 of the Staggers Rail Act of 1980, 49 U.S.C. § 10501(d), does not (contrary to the view of the Commission and Burlington) cut off district court jurisdiction with respect to shipments after the effective date of Staggers, October 1, 1980.3
We need not resolve the jurisdictional dispute, however. We held in Cross-Sound Ferry Services, Inc. v. ICC, 934 F.2d 327 (D.C.Cir.1991), as we had in prior cases, that when the merits of a case are clearly against a party seeking to invoke the court’s jurisdiction, and the jurisdictional question is exceptionally difficult, we may rule on the merits without addressing the jurisdictional issue. Id. at 333. Here, in the section of the case relating to multicar and trainload shipments, over which we have undisputed jurisdiction, we reject all the shippers’ attacks on the ICC’s application of R/VC, attacks on which McCarty had to prevail in order to succeed on the single-car shipments (of both wheat and barley). Thus, the decision on the segment where our jurisdiction is clear fully determines the outcome on the segment as to which it is disputed.
The context here differs from that of Cross-Sound Ferry and its predecessors in that McCarty, the party attacking our jurisdiction to give relief on single-car ship-[6]*6merits, is the same party that seeks relief against the agency (contingent on McCarty’s not winning the jurisdictional argument). But that does not give any call not to apply Cross-Sound Ferry. The unifying point is that, regardless of the outcome on the jurisdictional issue, the ultimate disposition is the same. In a key respect the ease is stronger for use of Cross-Sound Ferry than was that case itself; resolving the merits issues is exceptionally easy (at least in the sense that it requires no extra expenditure of judicial resources) because we necessarily must resolve them in the portion of the case over which we plainly have jurisdiction.4
II. The ICC’s Exclusion of Domestic Traffic
The ICC determined that the complaints before it challenged only the rates for shipments sent to certain Northwest Pacific ports for export. See McCarty Farms I, 4 I.C.C.2d at 264-66 (1988). The shippers assert that this interpretation was arbitrary and capricious. According to the shippers, the filings with the district court and the ICC consistently challenged not only those rates but also ones on traffic moving to interior points in Washington and Oregon, and traffic moving to ports for domestic consumption rather than export. We find no error in the Commission’s decision.
ICC rules of pleading do not require that origin and destination points be pled with specificity; the rules do mandate, however, that a complaint must “advise the Commission and the defendant fully” of alleged violations and provide “a detailed statement of the relief requested.” 49 CFR § 1111.1 (1990).
The Montana Department of Agriculture’s complaint as to Burlington’s trainload and multi-car rates was clearly limited to Portland and Seattle; it did not embrace all Pacific Northwest destinations, as is now claimed. See Complaint in No. 37815S (filed Mar. 26, 1981) at 6. However, the complaint did not clearly state whether only export traffic was being challenged. Nonetheless, the complaint included a verified statement of Eugene Radermacher that indicated that the complaint was limited to export rates. See McCarty Farms I, 4 I.C.C.2d at 264 n. 6.
The reasonableness of the ICC’s interpretation of the Montana Department of Agriculture’s complaint is underscored by the position taken by the Department and shippers in the proceedings on market dominance. There the complainants rested heavily on evidence that Montana wheat’s high percentage of protein and gluten made it uniquely attractive to purchasers in the Pacific Rim, thereby deflecting Burlington’s claim that competition from foreign sources would undermine any market power that Burlington might enjoy because of its virtual monopoly over carriage of Montana wheat to Pacific Coast export terminals. See McCarty Farms, Inc. v. Burlington Northern, Inc., 3 I.C.C.2d 822, 836-37 (1987). (The finding as to market dominance in barley is obscure. See id. at 837-39.) Thus the Commission’s finding of the reach of the Montana Department of Agriculture’s complaint was completely in accord with a key element of the complainants’ theory.
In interpreting McCarty’s complaint No. 37809 (which also governed in No. 37809 (Sub-No. 1), the challenge to Burlington’s single-car barley rates), the ICC noted that the complaint stated ambiguously that the destinations of the traffic at issue were “Pacific Coast terminals” in Washington and Oregon — a phrase that did not confine the issue to export terminals and grains. See McCarty I, 4 I.C.C.2d at 264 n. 6. Again, the Commission resolved the ambiguity by looking at McCarty’s accompanying submissions. Those included a verified statement of Robert L. Hines, Sr., which stated that the rates at issue were export rates, as well as subsequent testimony from McCarty witnesses Cecil Brennan and Robert L. Hines, who specified that the [7]*7dispute focused on rates for Montana’s export traffic in wheat and barley. See McCarty I, 4 I.C.C.2d at 264 n. 6 (citing Hines Verified Statement of 3/27/81 at 2). In addition, as with the Montana Department of Agriculture claims, the shippers’ analysis of market dominance matched the finding that they challenged only rates on export grain sent to the designated Pacific Coast ports. We do not find the Commission’s reading of the McCarty complaint arbitrary or capricious.
III. Propriety of the R/VC Methodology
Burlington’s sole claim is that the Commission, in determining reasonable rates, erred by dropping the Constrained Market Pricing (“CMP”) methodology that it adopted in the early 1980s in favor of another methodology (“R/VC”) that limits a railroad to charging rates for disputed traffic that yield no higher a ratio of revenue-to-variable-cost than the R/VC ratio of certain “benchmark” traffic used for comparison. First we trace the origin and theory behind the Commission’s adoption of CMP in the wake of the rail regulation reform legislation of 1976 and 1980. Then we consider the character of R/VC and the strength of the Commission’s explanation for adopting it in this case.
With the Railroad Revitalization and Regulatory Reform Act of 1976 (the “4R Act”) and the Staggers Rail Act of 1980, Congress moved to “allow[] the forces of the marketplace to regulate railroad rates wherever possible.” H.Conf.Rep. No. 1430, 96th Cong., 2d Sess. at 89 (Sept. 29, 1980) [U.S.Code Cong. & Admin.News pp. 3978, 4120, 4121], Recognizing that in many markets the growth of transportation alternatives had reduced the likely need for regulatory protection against rail monopoly, see S.Rep. No. 499, 94th Cong., 2d Sess. at 11 (Nov. 26, 1975), [U.S.Code Cong. & Admin.News pp. 14, 24, 25]. Congress in the 4R Act limited the ICC’s jurisdiction to situations where a carrier had “market dominance.” See Pub.L. No. 94-210, § 202(c)(i), 90 Stat. 35-36 (Feb. 5, 1976) (codified at 49 U.S.C. § 10709(c)).
In Staggers, Congress set a quantitative floor for the Commission’s jurisdiction over rates, and established that the Commission could not find market dominance if the disputed rate yielded an R/VC ratio lower than a specified figure — which started at 160% in the first year, rose in increments of 5% a year, and stabilized at 180% in the fifth. See 49 U.S.C. § 10709(d)(2)(A)-(E). Even where market dominance was shown, the Commission was to consider the railroads’ need to earn adequate revenue:
[i]n determining whether a rate established by a rail carrier is reasonable ... the Commission shall recognize the policy of this title that rail carriers shall earn adequate revenues, as established by the Commission under section 10704(a)(2) of this title.5
49 U.S.C. § 10701a(b)(3). Congress explained that the change was in response to the failure of its previous efforts to require the Commission to give attention to revenue adequacy:
Previous admonitions by the Congress that the Commission assist carriers in earning adequate revenue levels (49 U.S.C. § 10704) have not achieved their goals. As a result, the Committee is establishing a more straight forward mandate. This is a clear directive to ensure financially sound railroads, and the Commission is not to misuse the term “reasonable” to circumvent this directive.
H.Rep. No. 1035, 96th Cong., 2d Sess. at 54 (May 16, 1980) [U.S.Code Cong. & Admin.News p. 3999]. See also S.Rep. No. 499, 94th Cong., 2d Sess. at 11 (Nov. 26, 1975) [U.S.Code Cong. & Admin.News pp. 24, 25] (similar).
When the Commission began to develop its Coal Rate Guidelines in 1983, it adopted Ramsey pricing as “the appropriate analytical theory.” See Ex Parte No. 347 (Sub-No. 1), Coal Rate Guidelines— [8]*8Nationwide (unpublished decision issued Feb. 8, 1983), hereinafter 1983 Interim Coal Rate Guidelines at 9. It explained that Ramsey pricing was designed for cases where marginal cost is below average cost. Where that is true, a regulated firm forced to sell at marginal cost cannot recoup its total costs. Under Ramsey pricing, the. regulator allows firms to charge each user a premium over marginal cost in inverse proportion to the elasticity of the user’s demand. Because the highest charges fall on the most inelastic demanders, the impact on total usage is minimized. Thus, the Commission believed, it would reconcile the railroad’s need for revenue to cover total costs with the least possible distortion of demand (i.e., railroad usage would approximate as nearly as possible the level that would prevail under perfect competition). See id.; see also Ex Parte No. 347 (Sub-No. 1), Coal Rate Guidelines, Nationwide, 1 I.C.C.2d 520, 526 (1985).
Because of the difficulties in getting the data needed to apply Ramsey pricing in its pure form, the Commission adopted Constrained Market Pricing. Under CMP, the Commission allows a rail carrier to set prices freely in response to market forces, but provides some relief for captive shippers. Under its “revenue adequacy” approach, the Commission will focus on total system costs, corrected for “demonstrated management inefficiencies”, see id. at 534 & n. 35, with the understanding that costs not attributable to specific service may be recovered by differential pricing, i.e., by charges to captive shippers. Id. Regardless of what rate might emerge from this analysis, however, the final rate is subject to a cap based on stand-alone cost (“SAC”) — the full cost of providing the transportation for a shipper (or group of shippers) through a hypothetical, efficient, stand-alone rail operation that would serve only that shipper (or group of shippers). See id.; see also 1983 Interim Coal Rate Guidelines, at 11. With a SAC ceiling, no shipper (or shipper group) subsidizes others, at least in a strict sense of the term: though some bear a higher share of fixed costs than others, they still pay no more than what they would for a facility designed to serve only them. See id. at 11; Coal Rate Guidelines, Nationwide, 1 I.C.C.2d at 528. Further, in the Commission’s view SAC assures that even the most captive shippers pay no more than a “simulated competitive price.” Id. at 528, 542.
Although originally developed in the context of coal traffic, the Commission has consistently affirmed that CMP, with its SAC constraint, is the “preferred and most accurate procedure available for determining the reasonableness” of rates in markets where the rail carrier enjoys market dominance. See McCarty Farms, Inc. v. Burlington Northern, Inc., 3 I.C.C.2d 822, 839-40 (1987). Accord Ex Parte No. 347 (Sub-No. 2), Rate Guidelines — Non-Coal Proceedings, slip op. at 1-2 (served Apr. 8,. 1987) (“Constrained Market Pricing ... is the most sophisticated and accurate method available for determining the reasonableness of captive rates”).
In the McCarty rate challenges, however, the Commission applied the R/YC method rather than CMP/SAC. Under the R/VC method, a captive rate is deemed reasonable if its mark-up over variable cost is no greater than the mark-up on “benchmark” traffic selected as suitable for comparison. (Here, the Commission chose all western wheat and barley shipments of similar distance and with mark-ups higher than the ultimate jurisdictional threshold established by Staggers (180%). See McCarty Farms I, 4 I.C.C.2d at 275.) The Commission claimed that the R/VC analysis would “act[j as an indirect measure of Ramsey pricing” because “[b]y selecting traffic for the sample [the comparison traffic] with similar transportation and demand characteristics to the issue traffic, we are giving effect to the principles of Ramsey pricing and allowing differential pricing based on demand.” See McCarty Farms, 3 I.C.C.2d 822, 842 (1987). Thus the Commission purported to achieve something close to the economic rationality of CMP/SAC analysis while reducing the amount of information needed (and thus the cost of the calculation).
[9]*9Obviously the Commission is free to make reasonable trade-offs between the quality and cost of possible regulatory approaches. Reasonableness depends on how much quality is sacrificed for how much saving in cost, and of course we owe the Commission’s judgment on the point great deference. Here, however, the Commission has not intelligibly explained why the trade-off chosen was reasonable.
No one here attacks the Commission’s accolades for CMP/SAC. We therefore turn to the asserted drawbacks of R/VC, starting with a fundamental conceptual problem. If the formula is employed regularly and repeatedly, it will reduce rates to the lowest R/VC used in the comparison group. Suppose, for example, that the initial group (including the issue traffic) consisted of equal blocks of traffic with respective R/VCs of 180%, 185%, 190%, 195% and 200%. The average is clearly 190%, so that if the initial issue traffic comprises the blocks at 195% and 200%, they will be shaved to 190%. But at that point the average will have fallen (to 187%), and all the rates at 190% will be vulnerable to further reductions. These reductions in turn lower the average, and so on. Equilibrium would ultimately be achieved at 180%. Thus, having chosen 180% because Congress had identified it as the point at which the Commission should start to be concerned with possible abuse of market power, the Commission would find itself imposing 180% as a rate ceiling.
The shippers’ economics expert candidly acknowledged the truth of this difficulty, but answered it in practical terms. Because pre-Staggers rates that were not challenged within 180 days of its effective date could not thereafter be challenged, see Staggers § 229, 49 U.S.C. § 10701a note, the scenario of endless iteration was implausible. Verified Statement of Alfred Kahn, September 30,1987 at 8-9. The ICC adopted a similar argument. McCarty Farms I, 4 I.C.C.2d at 279. Further, of course, the process would require many iterations, and the relatively small gams for shippers from the later stages would dissuade them from bothering to drive the process forward. This practical defense, however, does nothing to give R/VC any glimmer of supporting principle or intellectual coherence.
Second, it is hard to discern any principle behind the choice of benchmark traffic. The Commission explained that it selected traffic with similar transportation and demand characteristics. McCarty Farms, 3 I.C.C.2d at 842. But if a profit-maximizing railroad charges more for the issue traffic than for the benchmark traffic, that would seem to represent a judgment — by the party with the greatest interest in making it correctly — that the issue traffic either costs more to transport or has a less elastic demand. If applied steadfastly, then, the R/VC method puts the Commission in the position of simply correcting (or purporting to correct) railroad management decisions as to what prices will most benefit the railroad. As it is hard to believe the Commission really thinks that would be a good allocation of its resources, it seems most probable that the transportation and demand characteristics of the two groups are in fact different. Thus, though the outcome of the method depends critically on the choice of benchmark, the Commission’s statement of its criteria of choice seems not to fit its purpose in the exercise, leaving the true criteria obscure.
Further, while CMP explicitly builds in the idea of revenue adequacy (subject to the SAC constraint), Coal Rate Guidelines, Nationwide, 1 I.C.C.2d at 534 & n. 35, the relation of R/VC to revenue adequacy is most obscure. The Commission observes (without dispute) that the rates it has set are above average cost and thus contribute to revenue adequacy. McCarty I, 4 I.C.C.2d at 280-81. But as Burlington’s revenues were inadequate — by the Commission’s own calculations — for each of the six years for which refunds were ordered,6 [10]*10it is not altogether clear why the rates charged constituted an abuse of market power within the Staggers Act frame of reference, with its heavy stress on revenue adequacy. This obscurity is underscored by the haphazard character of the choice of benchmark traffic.
Thus, while the Commission urges that R/VC is really a variant of Ramsey pricing, this is so only in the vaguest imaginable sense, namely, that it leaves the railroad's lower rates alone. The principle for limiting the higher rates has no evident connection to any of the goals that the Commission said CMP/SAC was designed — indeed, well designed — to achieve.
It is hardly surprising, then, that the shippers' own economics expert conceded serious drawbacks in R/VC, candidly admitting that it “provides no economic principle, no theoretically definable test, against which the ‘reasonableness’ of the markups [over the variable cost of captive rates] can in principle be objectively” assessed. Verified Statement of Alfred Kahn, September 30, 1987 at 4. “All it [R/VC analysis] ... accomplishes] ... [is] to ensure that no single shipper or group of shippers ... is exploited more than all other captive shippers.” Id. at 5. Even the last comment does not quite capture the problem, for it assumes exploitation, which R/VC principles do not intelligibly define.
Perhaps there are things to be said in favor of R/VC methodology that the Commission has left unsaid; of course it may seek to fill the gap on remand. The present record, however, suggests a profound discrepancy in the quality of the competing methods. That takes us to the Commission’s cost justification. The Commission claims that when, as here, there are multiple, scattered points of origin, it is very costly (or perhaps even impossible) to develop stand-alone figures for rates under review. Thus it argues that while CMP/ SAC pricing may be best for coal pricing, coal carriage is unique because coal “is typically shipped in frequent, regular, high-volume (unit-train) movements over high-density rail lines between a very few origin and destination points.” See Commission Brief at 26.
This argument is not supported by the ICC’s own treatment of apparently analogous cases. In Metropolitan Edison Co. v. Conrail, 5 I.C.C.2d 385 (1989), the Commission used CMP/SAC pricing to evaluate the rates of coal shipments that seem quite similar, in just these respects, to the disputed wheat and barley shipments here. In Metropolitan Edison, there were 39 different origination points, several low-density routes, and a total annual volume of coal at issue — 440,000 tons — that was much lower than the 2,000,000-3,500,000 tons/year of grain at issue in these proceedings. See id. at 416, 419-20, 423; Burlington Comments at 14-15 (filed August 7, 1987). Moreover, the Metropolitan Edison decision never suggested that it was particularly hard to develop SAC data for the type of traffic under review.
Similarly, in Coal Trading Corp. v. Baltimore & Ohio R.R., 6 I.C.C.2d 361 (1990), the ICC analyzed rates with the CMP/SAC methodology, notwithstanding that the case involved 507 origin-destination pairs, with shipments from five states terminating at three coal piers on the coasts of Maryland and Virginia. See id. at 362. That situation appears quite similar to the present one, in which grain moved from 272 origination points in Montana to a handful of destination points on the Pacific Northwest Coast. See Verified Statement of Philip H. Burris, August 6, 1987 at 3. For origin-destination pairs with very few shipments, the Commission also appeared to believe that some sort of extrapolation could be used. See Coal Trading Corp, 6 I.C.C.2d at 377 n. 15. Thus the Commission has by no means shown that the number of origin-destination pairs or the density of the traffic raise the information costs so drastically as to justify scuttling CMP.
Burlington submitted SAC calculations in this case in an effort to demonstrate their [11]*11feasibility. The Commission critiqued the data on grounds that the assumptions were unduly self-serving and unreliable, showing Burlington had “little incentive to develop [data for] a least cost [SAC] system.” See McCarty Farms I, 4 I.C.C.2d at 271; see also Commission Brief at 27-28 (attacking Burlington for its “self-serving assumptions”). This is hardly a smashing riposte. Of course no adjudicator would expect to be able to rely entirely on one side’s analysis. We hope that the Commission does not think that the CMP/SAC model works for coal because parties in that type of litigation never make self-serving assumptions. Burlington’s submissions showed that a SAC analysis could be done; as to whether the shippers and the Commission could not correct the defects at a reasonable cost, the Commission never spoke.7
In McCarty Farms I the Commission suggested that the costs of CMP posed a special problem where the claim was brought by small shippers who would not be able to recover their expenses except in the event of victory, 4 I.C.C.2d at 267-68, but the Commission does not press the point in its brief here. The concern seems quite weak where, as here, the shippers’ claims are pressed by an entity that in a sense aggregates all shipper interests — the Montana Department of Agriculture. Even without the state presence, one can imagine risk-allocating devices, such as allowing a successful representative shipper to recover double its reasonable costs out of the total overcharges collected, which would provide adequate incentives for the additional investment — if indeed CMP is materially more costly to apply than R/VC.
Given the Commission’s praise for CMP/ SAC, it can on remand support its switch to R/VC only by a careful analysis indicating that the incremental costs of applying CMP/SAC are large in relation to the attendant sacrifice of quality (which, of course, it may seek to deprecate). On the present record, however, the jettisoning of CMP/SAC cannot pass for reasoned decisionmaking.
IV. Shippers’ Attacks on the Application of R/VC
We here address a series of criticisms of the way the Commission went about applying the R/VC methodology. The shippers’ attacks on that application are now ripe and fully briefed, and we cannot completely discount the possibility that the Commission on remand will lawfully adhere to the R/VC methodology. Thus the most practical course, in terms of economy of judicial and litigants’ resources, and the interest in speeding the controversy to its end, is to dispose of the claims.
The lack of any visible intellectual coherence in the R/VC method, noted above, somewhat colors this process. The shippers are in the position of one who, having insisted that a dart-thrower throw blindfolded and with his back to the dartboard, now claims that the dart-thrower would do better if he added a little twist of the wrist.
A. ICC’s Choice of Jurisdictional Thresholds
The first claim stems from the way the Commission used the jurisdictional thresholds established by the Staggers Act. Recall that the Act set up a progressive limitation on the Commission’s jurisdiction, barring it from reviewing the reasonableness of rates below certain R/VC ratios. The threshold ratio was 160% for the first year (October 1, 1980 through September 30, 1981), and rose by 5% a year until stabilizing, effective October 1,1984, at the present ratio of 180%. In defining the benchmark rates, the Commission excluded all traffic with rates yielding R/VC ratios below 180%. The shippers had argued that the benchmark traffic should reach lower— to all traffic that more than covered the railroad system’s average costs. See Veri[12]*12fied Statement of Alfred Kahn, September 30, 1987 at 5.
The shippers invoke a variety of arguments. First, they argue that the Commission applied the jurisdictional threshold in a mistaken belief that it had no legal authority to include traffic yielding lower revenue. This is simply a misreading of the Commission, which clearly understood itself to be making a policy judgment. See McCarty Farms I, 4 I.C.C.2d at 275-76 & n. 26; McCarty Farms, Inc. v. Burlington Northern, Inc., Docket Nos. 37809, 37809 (Sub-No. 1), 37815S (unpublished decision served May 11, 1988) {“May 1988 Decision ”) at 3.
Second, the shippers argue that by excluding traffic with revenue below the 180% R/VC ratio but above fully allocated costs, the Commission wrongly “exclude[d] any competitive traffic.” McCarty Brief at 29. But the Commission’s explanation seems quite understandable. Its aim in applying the R/VC test, it said, was to “develop a benchmark of maximum reasonableness based on the differential pricing experienced by similarly situated traffic, not on a ratio which represents an average mark-up.” McCarty Farms I, 4 I.C.C.2d at 275-76 (emphasis in original). This appears completely in accord with the undisputed proposition that Staggers contemplated that rates could be reasonable even though they reflected an exercise of market power. See 49 U.S.C. § 10709(c) (“a finding of market dominance does not establish a presumption that the proposed [or existing] rate exceeds a reasonable maximum”). We note that the shippers’ own economics expert said the he found “no basis in principle for choosing between the Commission’s and [the shippers’] proposals.” Verified Statement of Alfred Kahn, September 30, 1987 at 9. Further, the shippers’ proposal incorporates a principle which, if it were continuously applied, would reduce the rates paid by captive shippers to their share of fully allocated cost. Compare p. [16] above. Since railroads cannot expect to recover the very elastic shippers’ share of fully allocated costs from them, the shippers’ principle would condemn the railroads to revenue inadequacy.
Third, the shippers assert that if the Commission was going to use the jurisdictional threshold as a floor in defining the benchmark rates, it should have used, for each year, the threshold applicable to that year, e.g., 160% for October 1, 1980 through September 30, 1981. The Commission, however, adequately explained its reasoning in rejecting this idea. It viewed Congress’s incremental lifting of the jurisdictional floor as an effort “primarily to cushion the potential of rate increases in the wake of loosened federal controls.” May 1988 Decision at 3 n. 5. In the ICC’s view, the successive floors spoke not to what was a reasonable rate, but only to what rates might justify Commission concern. Id. at 2. Because “the question [of benchmark rates] relates to an approximation of the point of [market] captivity, not the impact of sudden increases,” it made sense to choose “[t]he end point” of the alteration in jurisdictional threshold. Id. at 3 n. 5. Given the necessarily unscientific character of the exercise as a whole, the judgment seems by no means arbitrary.
B. The Commission’s Failure to Use Movement-Specific Costs
The shippers object to the Commission’s decision to use generic system costs for wheat and barley shipments instead of movement-specific costs. They argue that the Commission has previously used movement-specific costs, and note that the Commission has endorsed that approach as “sound in concept” under “normal circumstances”. McCarty Farms I, 4 I.C.C.2d at 282.
The shippers’ objection is unpersuasive. First, the Commission employed system average costs when computing variable costs for the benchmark traffic, evidently without objection by the shippers. See McCarty Farms II, at 4; McCarty Farms I, 4 I.C.C.2d at 282. Noting that the benchmark and issue traffic were very similar, the Commission believed that any adjustment of costs for the issue traffic would in all likelihood be suitable for the comparison [13]*13traffic as well. Thus, if the adjustment were made on both sides, it might well be pointless; if on only one, it would create phony discrepancies. Id. Further, the shippers themselves had offered cost adjustments only for two years, using extrapolation for the rest. Id. The shippers’ suggestion would have dragged out the proceedings, with no apparent improvement in accuracy.
C. Use of the R/VC Standard as a Cap for Average Rather Than Individual Shipment Rates
The shippers complain that the Commission acted arbitrarily and capriciously when it applied the R/VC standards as a cap which the average of Burlington’s rates (as opposed to each single rate) could not exceed. More specifically, when applying the R/VC standard, the ICC grouped the issue traffic by year and then additionally by type (trainload, multiple-car, and single car); it then offset rates producing an R/VC ratio lower than the permissible maximum against those rates with an above-standard R/VC ratio. See McCarty Farms I, 4 I.C.C.2d at 278. The shippers say that this application of the R/VC standard conflicts with Commission precedent and fails to ensure that every individual shipper is protected against excessive rates.
The Commission’s answer seems persuasive. If .all Burlington’s above-standard rates were lowered to the standard, when Burlington could not retroactively collect for below-standard charges, the railroad in the aggregate would receive less than the standard. Id. The Commission’s choice here seems a logical response to the shippers’ decision to have advanced their claims as a class action, seeking relief on an aggregate rather than shipment-by-shipment basis.8
D. Reliance on Indexing
The Commission calculated the permissible R/VC for 1979-86 with aggregate data from those years; for later years, it took the maximum rates for 1986 (derived from the aggregate benchmark) and indexed them forward with a “Rail Cost Adjustment Factor.”9 The shippers argue that when 1987 data became available, the Commission should have recalculated the R/VC benchmarks with the 1987 data. We see no need for such recalculation. If sorting out disputes about the new data took much time (and there were disputes over the 1987 data, see McCarty Farms III, slip op. at 3), any such requirement could drag proceedings out indefinitely, as the digestion of one year’s data allowed the next year's to emerge, itself ready for digestion. Cf. Cost Ratio for Recyclables — 1983 Determination, Ex Parte No. 394 (Sub-No. 1), 3 I.C.C.2d 407, 410 (1985) (noting impossibility of using completely current information).
There is quite an irony in the shippers’ harping upon the Commission’s having taken shortcuts in its application of R/VC, when the latter methodology can itself be justified (if at all) only as an economy of effort. But whereas Burlington has shown a serious qualitative disadvantage in R/VC as opposed to CMP/SAC, the shippers seem to insist on endless increases in data analysis by the Commission in the application of R/VC, without suggesting any compelling gains in accuracy or conceptual coherence.
[14]*14E. Method of Calculating Interest
The shippers argue that the ICC acted “arbitrarily, capriciously and unreasonably” by not employing its usual method of calculating interest. See 49 CFR § 1141.1.10 When an agency regulation is invoked, the first step of analysis, of course, is to determine whether it is applicable by its terms, a matter on which we owe the agency great deference. See Sandstone Resources, Inc. v. FERC, 973 F.2d 956, 959 (D.C.Cir.1992).
As the Commission has explained, the ICC’s rules have no provisions for class action complaints. As a consequence, when the district court referred the McCarty suit, the Commission was confronted with a novel problem. 49 CFR § 1141.1 contemplates selecting an interest rate based on rates prevailing on the date of the first unlawful shipment. See McCarty Farms III, slip op. at 4. But here, rather than assessing the reasonableness of charges for specific shipments, the Commission calculated the reasonableness of various categories of rates taken on an aggregate and annual basis. See McCarty I, 4 I.C.C.2d at 278-79. The damage awards in this case therefore did not represent an actual overcharge owed to any specific individual, but instead an aggregated overcharge for each year owed to multiple unidentified shippers. Literal application of § 1141.1 therefore made little sense. The Commission’s decision to use a single interest rate for each year, and to assume that there was one overcharge occurring in the middle of each year, see McCarty Farms III, slip op. at 4-5, was anything but arbitrary.
F. Prospective Rate-Setting
McCarty says that the ICC erred by refusing to prescribe prospective rate ceilings for all the rates challenged in these proceedings, thereby depriving “the Montana shippers ... [of] the protection afforded them by the Staggers Act.” McCarty Reply Brief at 18.
Once the Commission finds a rate in violation of the statute (but only then), it “may prescribe the rate ... to be followed” in the future. 49 U.S.C. § 10704(a)(1) (emphasis added). Concerning single-car and multiple-car (26-car) shipments, therefore, as to which the Commission had found no unreasonable rates, it lacked authority to prescribe future ones. See McCarty Farms IV, slip op. at 3-4. Accord Central of Ga. R.R. Co. v. United States, 379 F.Supp. 976, 979 (D.D.C.1974), aff'd sub nom. United States Clay Producers Traffic Ass’n v. Central of Ga. R.R., 421 U.S. 957, 95 S.Ct. 1944, 44 L.Ed.2d 446 (1975).
For trainload rates, the Commission originally prescribed prospective rate ceilings; it then elected to lift them after concluding that they were unnecessary. See McCarty Farms IV, slip op. at 4. In so doing, it relied on data which showed that levels for trainload rates had been reasonable from 1987-90, noted findings that the export grain market was now competitive, drew on its experience to conclude that the situation was unlikely to change in the future, and noted that shippers could bring complaints to the ICC if they regarded future rates as unreasonable. See id. at 3-4. There was nothing arbitrary there.
******
On the Montana Department of Agriculture component of the dispute, we remand the case to the ICC to reconsider whether it was appropriate to use R/VC instead of CMP/SAC pricing to evaluate the reasonableness of the multi-car and trainload rates; we also reject the shippers’ claims [15]*15as to the application of R/VC. On the basis of these determinations and Cross-Sound Ferry, we reject the McCarty claims as to alleged error in the Commission’s treatment of the subjects of disputed jurisdiction, namely, all single-car shipments of barley, and all disputed wheat shipments after September 12, 1980.
So ordered.