Opinion for the Court filed by Senior Circuit Judge HARRY T. EDWARDS.
HARRY T. EDWARDS, Senior Circuit Judge.In 1991, Congress instructed the Department of Transportation (“DOT”) to determine whether drivers of commercial motor vehicles (“CMVs”) — large trucks, passenger coaches, and school buses— were receiving adequate training. Inter-modal Surface Transportation Efficiency Act, Pub.L. No. 102-240, 105 Stat.1914, 2151 (1991) (“ISTEA” or the “Act”). In July 1995, after extensive study, the Federal Highway Administration (“FHWA”) published a three-volume study entitled, “Assessing the Adequacy of Commercial Motor Vehicle Driver Training: Final Report” (“Adequacy Report”). The Report concluded, inter alia, that in order for any training program to be “adequate,” it must include “on-street hours” of training. The findings of the Adequacy Report were distilled into a Final Regulatory Evaluation, which the agency transmitted to Congress in February 1996. In April 1996, the agency published a notice in which it solicited comments on the Adequacy Report and the Final Regulatory Evaluation. And then nothing much happened until November 2002, when parties petitioned this court for a writ of mandamus ordering the Secretary of Transportation to fulfill his ISTEA duties. In re Citizens for Reliable & Safe Highiuays, No. 02-1363 (D.C.Cir.Nov. 26, 2002). Pursuant to a settlement agreement, the agency agreed to publish a final rule implementing entry-level training requirements no later than May 31, 2004.
On August 15, 2003, the Federal Motor Carrier Safety Administration (“FMCSA”) published a notice of proposed rulemaking to address the findings of the Adequacy Report. After eliciting comments, FMCSA issued a final rule in May 2004. In the rule’s summary, FMCSA stated: “This action responds to a study mandated by the Intermodal Surface Transportation Efficiency Act of 1991 that found the private sector training of entry-level drivers in the heavy truck, motorcoach, and school bus industries was inadequate.” Minimum Training Requirements for Entry-Level Commercial Motor Vehicle Operators, 69 Fed.Reg. 29,384, 29,384 (May 21, 2004) (codified at 49 C.F.R. pt. 380). However, the rule departed sharply from earlier agency recommendations. The Adequacy Report determined that effective training for CMV drivers required practical, on-the-road instruction on how to operate a heavy vehicle. But FMCSA ignored this evidence and opted for a program that focuses on areas unrelated to the practical demands of operating a commercial motor vehicle.
Petitioners, who represent private citizens concerned with highway safety and the industries affected by training requirements, seek review of FMCSA’s final rule. The striking incongruity between the methods of training previously shown to be effective and the regimen adopted in the final rule, petitioners argue, shows the agency’s action to be arbitrary and capricious, in violation of the Administrative Procedure Act (the “APA”). See 5 U.S.C. § 706(2)(A). We agree. Initial phases of the regulatory process — which involved extensive study and voluminous reports— identified deficiencies in training, and then prescribed standards for judging training *339adequacy and guidelines for inculcating the requisite operational skills. FMCSA proclaims that its final rule “responds” to those earlier findings. In truth, however, the final rule inexplicably ignores the Adequacy Report and the regulatory prescriptions contained in that Report. The agency, without coherent explanation, has promulgated a rule that is so at odds with the record assembled by DOT that the action cannot stand. Accordingly, we grant the petitions for review and remand the final rule to the agency for further rulemaking consistent with this opinion.
i. Background
A. Licensing and Training Drivers of Commercial Motor Vehicles
This case concerns Congress’s ongoing efforts to ensure that CMVs operate safely on the nation’s roads. For almost two decades, the federal government has regulated the licensing of CMV drivers. However, prior to the instant rulemaking, which was instituted under ISTEA, the Government never purported to impose any standards of driver training. Private parties had developed training for neophyte drivers, but these efforts were found to be insufficient to secure CMV safety.
In 1986, Congress passed the Commercial Motor Vehicle Safety Act (“CMVSA”), 49 U.S.C. § 31301 et seq. (2000). Under the CMVSA, the Secretary of Transportation was required to promulgate regulations, to be administered by individual states, setting minimum uniform standards governing commercial drivers’ licenses (“CDLs”) for CMVs. Id. § 31308. CMVs include cargo-carrying trucks within a specified weight range, vehicles designed to transport at least 16 passengers, and vehicles carrying certain hazardous materials. Id. § 31301(4). Among other things, the statute mandates that CDL tests include written and driving components. Id. § 31308(1).
The federal standards governing CDLs do not establish a training regimen. In other words, “there are no prerequisite Federal training requirements to obtain a CDL.” Minimum Training Requirements for Entry-Level Commercial Motor Vehicle Operators, 68 Fed.Reg. 48,863, 48,864 (proposed Aug. 15, 2003) (codified as amended at 49 C.F.R. pt. 380). “Generally, drivers individually prepare for the CDL tests by studying such areas as vehicle inspection procedures, off-road vehicle maneuvers and operating a CMV in traffic.” Id.
While the CDL program does not mandate any CMV training, some segments of the private sector, with guidance from the federal government, have attempted to promote effective training practices. In 1985, FHWA published a Model Curriculum for Training Tractor-Trailer Drivers (“Model Curriculum”). See Joint Appendix (“JA.”) 37. The Model Curriculum sets out a primer for instructing drivers of heavy trucks. It focuses on five subject areas: basic operation, safe operating practices, advanced operating practices, vehicle maintenance, and nonvehicle activity. The Model .Curriculum prescribes a total of 320 hours of training, including 116 hours of on-street training and 92.25 additional hours of driving-range time. Id at 44. The curriculum is primarily focused on inculcating the skills and knowledge needed to enhance CMV safety. For example, it prescribes 4.25 hours of training on the techniques needed to avoid accidents while driving a truck in reverse, and 22 hours on “advanced operating practices,” like emergency maneuvers and skid control. Id. Still, the Model Curriculum’s introduction emphasizes that its program sets out only “minimum standards,” and that “[graduates of this Curriculum cannot be considered fully trained, .‘ready to solo’ type drivers” unless “the Curriculum is considerably expanded and enriched to *340provide both additional driving time and material pertinent to the particular driving job that the student is being trained for.” Id. at 42 (emphases in original). In 1995, FHWA devised a similar curriculum for motor coach drivers.
Shortly after the Model Curriculum was published, groups representing the motor carrier, truck-driver training, and insurance industries formed the Professional Truck-Driver Training Institute of America (“PTDIA” or the “Institute”). The Institute develops standards for training truck drivers, and it certifies private training organizations that meet or exceed its recommendations. PTDIA acknowledges that the Model Curriculum “has been the ‘bible’ around which the PTDIA has built its standards.” Professional Truck Driver Institute of America, Comments to 49 C.F.R. pt. 383, at 3, reprinted in J.A. 68. To qualify as adequate under PTDIA standards, a track driver training program must provide 147.5 hours of instruction including 44 hours of combined street and range time. Id. at 10, J.A. 75.
Congress revisited the issue of CMV safety in 1991 when it passed ISTEA. The rulemaking at issue here was commenced pursuant to § 4007(a) of the Act, which provides:
(a) ENTRY LEVEL.-
(1) STUDY OF PRIVATE SECTOR. — Not later than 12 months after the date of the enactment of this Act, the Secretary shall report to Congress on the effectiveness of the efforts of the private sector to ensure adequate training of entry level drivers of commercial motor vehicles. In preparing the report, the Secretary shall solicit the views of interested persons.
(2) RULEMAKING PROCEEDING. — Not later than 12 months after the date of the enactment of this Act, the Secretary shall commence a rule-making proceeding on the need to require training of all entry level drivers of commercial motor vehicles. Such rulemaking proceeding shall be completed not later than 24 months after the date of such enactment.
(3)FOLLOWUP STUDY. — If the Secretary determines under the proceeding conducted under paragraph (2) that it is not in the public interest to issue a rule that requires training for all entry level drivers, the Secretary shall submit to the Committee on Commerce, Science, and Transportation of the Senate and the Committee on Public Works and Transportation of the House of Representatives not later than 25 months after the date of the enactment of this Act a report on the reasons for such decision, together with the results of a cost benefit analysis which the Secretary shall conduct with respect to such proceeding.
105 Stat. at 2151-52.
B. Implementing ISTEA’s Dictates
1. Studying the Adequacy of Entry-Level Driver Training
As a first step toward implementing § 4007(a) of ISTEA, FMCSA’s predecessor, FHWA, issued an advanced notice of proposed rulemaking. Training for All Entry Level Drivers of Commercial Motor Vehicles (CMVs), 58 Fed.Reg. 33,874 (announced June 21, 1993) (to be codified at 49 C.F.R. pt. 383). Noting the requirements of § 4007(a), FHWA solicited comments on “the need to require training of all entry level drivers of commercial motor vehicles (CMVs).” Id. The agency further explained that it had contracted with an outside company to produce a study examining the effectiveness of private sector efforts to ensure adequate training of entry-level CMV drivers. Id. at 33,875. Information gleaned from this study would form the basis of the report to Congress *341mandated by ISTEA. To help assemble the necessary material, the agency invited interested parties to submit comments in response to a list of 13 questions addressing the state of entry-level driver training. Id. at 33,875-76.
In July 1995, after receiving over 100 responses to its request for comments, FHWA published its three-volume Adequacy Report, reprinted in J.A. 164. The Adequacy Report began by surveying the training levels among drivers of heavy trucks, motor coaches, and school buses. What it found was not encouraging: “The conclusion of this study is that none of the three private sectors are effectively providing adequate training.” 1 Adequacy Report at 2, J.A. 173. Specifically, the Adequacy Report found that “the heavy truck sector has the smallest percentage of carriers offering adequate training (about 9 percent),” while only 18.5 percent of motor coach carriers offered adequate training. Id. at 3-4, J.A. 174-75. The Report concluded that “the present level of training adequacy is not likely to improve due to the actions of the private sectors themselves.” Id. at 7, J.A. 178. Given widespread training failures across the industries it examined, the Report recommended that “[i]f it is desirable to target fewer than all three domains, the heavy truck domain should be considered first priority, followed by motorcoaehes.” Id. at 12, J.A. 183.
The Adequacy Report also made extensive findings on the form that “adequate” entry-level training would take. “With regard to heavy trucks,” the Report stated, “there is general agreement in the industry that the model tractor-trailer driver curriculum developed by the FHWA in the mid-1980s represents an adequate content and approach for training truck drivers.” 3 Adequacy Report at 1-6, J.A. 209. “Therefore, the model curriculum was the starting point in defining ‘adequate training’ for heavy truck drivers.” Id. Using the Model Curriculum as a baseline for analysis, the Report noted that “[f]or a program to be considered ‘adequate’-, it must have on-street hours.” Id. at B-5, J.A. 216.
The findings of the Adequacy Report were distilled into a Final Regulatory Evaluation, which FHWA transmitted to Congress in February 1996. The evaluation presented a cost-benefit analysis of mandating entry-level dxiver training in conformity with the Model Curriculum. On balance, the x'eport found, mandatory training would be beneficial. In its analysis, the agency discounted the anomalous results produced by some earlier studies of driver training, which had suggested that training might increase accident rates. Such “counterintuitive” findings, the ageixcy determined, likely reflected the pervasiveness of training programs that were “not adequate.” Final Regulatory Evaluation: Entry-Level Driver Training at 15-17 (May 1995), reprinted in J.A. 241-43. Adjusting for shoddy training programs made the benefits of good training clear. “Based on the information presented from case studies, a reduction in accidents is possible when training is well designed. Accident reductions in the 10 to 15 percent range are not unrealistic.” Id. at 20, J.A. 246. Economically, mandating training along the lines described in the Model Cuiriculum would yield substantial projected benefits. Against a cost of between $4.19 billion and $4.51 billion over 10 years, mandatox-y training was expected to generate a benefit in the range of $5.4 billion to $15.27 billion during the same period. Id. at 32-36, J.A. 258-62.
In Apx’il 1996, FHWA published a notice in which it solicited comments on the Adequacy Report and the Final Regulatory Evaluation. Training of Entry-Level Drivers of Commercial Motor Vehicles, 61 *342Fed.Reg. 18,355 (Apr. 25, 1996) (notice of availability and request for comments). In response, the agency received 48 additional comments. On November 13, 1996, the agency held a public meeting on the issue, which 26 individuals attended. After the meeting, however, the agency’s activities pursuant to § 4007(a) came to a halt. Nothing in the record explains this hiatus, but for six years the agency initiated no further action. Indeed, it took litigation by concerned private parties to nudge the agency out of its slumber. In November 2002, those parties petitioned this court for a writ of mandamus ordering the Secretary of Transportation to fulfill his ISTEA duties. In re Citizens for Reliable & Safe Highways, No. 02-1363 (D.C.Cir.Nov. 26, 2002). The matter was settled, and DOT agreed to publish a final rule implementing entry-level training requirements no later than May 31, 2004. Id.
2. FMCSA’s Final Rule
On August 15, 2003, FMCSA published a notice of proposed rulemaking. After recounting the findings of the Adequacy Report, FMCSA proposed a novel approach to the problem of CMV training:
The agency is not requiring entry-level drivers to receive training in areas that are covered in the CDL test. Such training would be redundant. Instead, the required training would address: (1) driver qualifications — medical, and drug and alcohol testing, (2) driver hours of service limitations, (3) driver wellness, and (4) whistle blower protection.
Minimum Training Requirements, 68 Fed. Reg. at 48,868. FMCSA estimated that training in this course of study would entail about 10.5 hours for heavy truck and motor coach drivers. Id. The proposed rule defined “entry-level” driver as one with less than two years experience. Id. at 48,869. Licensed drivers with one year of experience and a good driving record, however, would be “grandfathered” past the new standard. Id. Finally, the agency asserted that the proposed rule would be cost-justified, but it offered no studies directly demonstrating the rule’s economic benefit. Instead, FMCSA relied on inferences from data related to more extensive training regimens. Estimating a 10-year cost of $173.3 million, the agency claimed that the proposed rule would have to prevent 315 truck-related accidents in the first year and 285 crashes in subsequent years to be cost-beneficial. Id. Since those numbers represented less than one percent of truck-related accidents, and the training program contemplated in the Adequacy Report was projected to cut accident rates by up to 15 percent, the agency concluded that the scaled-down reduction in accidents required to justify the rule’s cost was attainable. Id.
After eliciting comments and holding a public meeting, FMCSA announced its final rule in May 2004. In the rule’s summary, the agency stated: “This action responds to a study mandated by the Intermodal Surface Transportation Efficiency Act of 1991 that found the private sector training of entry-level drivers in the heavy truck, motorcoach, and school bus industries was inadequate.” Minimum Training Requirements, 69 Fed. Reg. at 29,384. The final rule mandates action in the four areas sketched in the notice of proposed rulemaking. 49 C.F.R. § 380.503. FMCSA did not set a minimum number of hours, but suggested that 10 hours of training would be appropriate for heavy truck drivers. 69 Fed.Reg. at 29,398. The new rule defined “entry-level” to cover all drivers with less than one year of experience driving with a CDL. 49 C.F.R. § 380.502(b).
In July 2004, petitioners, Advocates for Highway and Auto Safety (“Advocates”) and the Owner-Operated Independent Drivers Association, "filed separate peti*343tions for review of FMCSA’s final rule. Petitioner United Motorcoach Association (“UMA”) initially filed a petition for reconsideration with FMCSA. It then filed a petition for review with this court, which was dismissed as incurably premature since UMA had a petition for reconsideration pending before the agency. After notifying the agency that it was withdrawing its petition for reconsideration, UMA filed a new petition for review with this court in December 2004. All three petitions were consolidated for review.
In 1996, Congress passed an Act to “codify without substantive change laws related to transportation and to improve the United States Code.” Pub.L. No. 104-287, 110 Stat. 3388 (1996). , Oddly, the Act seems to repeal § 4007(a). Id. § 7(8), 110 Stat. at 3400. Neither party mentions this Act, and it is hard to know what to make of it. The Act was passed in October 1996, approximately eight months after the FHWA fulfilled its mandate under § 4007(a)(1) by transmitting the Adequacy Report to Congress. Congress appears to have believed it was repealing only obsolete statutory language relating to the § 4007(a)(1) mandate. See h.R. Rep. 104-573, 23, 1996 U.S.C.C.A.N. 3831, 3853 (“Section 7(8) repeals section 4007(a), (c), (d), and (e) of the Intermodal Surface Transportation Efficiency Act of 1991 ... to eliminate obsolete provisions.”). However, beyond the obsolete requirements of § 4007(a)(1), the remainder of § 4007(a) unequivocally required the agency to take further action, either by issuing a rule on driver training or, if doing so would contravene the public interest, by submitting a further report to Congress. Id. § 4007(a)(2)-(3). Neither party suggests that Pub.L. No. 104-287 repealed the agency’s obligations beyond those set forth in § 4007(a)(1). Indeed, both sides argued the case as if the disputed statutory provisions remain in full force and effect.
Whatever the status of § 4007(a)— and for purposes of this opinion, we share the parties’ evident assumption that it remains operative — it seems clear that this court has authority to determine whether FMCSA’s final rule is arbitrary and capricious under the APA, 5 U.S.C. § 706(2)(A). FMCSA, whose chief mission is to ensure highway safety, see 49 U.S.C. § 113(b) (2000), has the undisputed authority to promulgate regulations responding to the findings of the Adequacy Report, see id. § 113(f) (authorizing the Secretary of Transportation to delegate authority to FMCSA). That is precisely what the agency set out to do when it issued its Notice of Proposed Rulemaking. See 68 Fed.Reg. at 48,863. And the parties do not doubt that the agency’s final regulations are subject to judicial review under the APA. In other words, for purposes of judicial review, it does not matter whether FMCSA’s final rule is viewed as an act taken pursuant to a specific duty under ISTEA or an act taken pursuant to the authority granted under the agency’s organic statute to address matters relating to highway safety. In either case, FMCSA lawfully set out to promulgate regulations that “respond” to the Adequacy Report. That is the basis upon which we review the final rule to determine whether it survives judicial scrutiny under the APA’s arbitrary-and-caprieious standard of review.
ii. Discussion
A. Standard of Review
A party challenging an agency’s rulemaking has the burden of showing that the agency action was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A); see also City of Olmsted Falls v. FAA, 292 F.3d 261, 271 (D.C.Cir.2002). An agency’s rule will be found arbitrary *344and capricious “if the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to difference in view or the product of agency expertise.” Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 468 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983).
B. FMCSA’s Entry-Level Driver Training Rule
The contested final rule begins with the assurance that it “responds” to the Adequacy Report. See Minimum Training Requirements, 69 Fed.Reg. at 29,384. Rather than “respond” to the imperatives laid out in the Adequacy Report, however, the final rule completely ignores the study’s emphasis on practical, on-the-road training. The agency has adopted a rule with little apparent connection to the inadequacies it purports to address. For this reason, it fails review under § 706(2)(A).
The record reveals numerous disjunctions between the final rule and the findings of the Adequacy Report. Most glaringly, the final rule inexplicably abandons the recommendations of the Model Curriculum, despite the Adequacy Report’s heavy reliance on those recommendations. The Report, in its definition of adequate training, says that “the model curriculum was the starting point in defining ‘adequate training’ for heavy truck drivers,” and that it is largely applicable to the motor coach industry. 3 Adequacy Report at 1-6, J.A. 209 (emphasis added). FMCSA accepted this premise when it announced the rule at issue here. Minimum Training Requirements, 68 Fed.Reg. at 48,865 (“The agency believes that the Model Curriculum represents the basis for training adequacy.”). Methodologically, the Adequacy Report is entirely structured around the notion that “adequate training” is defined in reference to the Model Curriculum. But the final rule eschews the Model Curriculum altogether. While the curriculum devotes some attention to the topics covered by the final rule — for example, it prescribes five hours of instruction on personal health and safety, and 5.75 hours of training on hours of service requirements — those subjects are clearly secondary. Model Curriculum at 3, reprinted in J.A. 44. Overwhelmingly, the curriculum addresses topics directly related to driving skills, with a heavy emphasis on skills and techniques necessary to safely operate a heavy truck. None of the four areas covered by the final rule — driver qualification, hours of service, driver wellness, and whistleblower protection— have anything to do with operational skills. Thus, they fly in the face of the Adequacy Report’s recommendations.
A critical facet of the training program developed in the Model Curriculum is on-street training. Indeed, the Adequacy Report’s conclusions evince a presumption that any rule instituting mandatory training would contain a substantial on-street training component. Early on, the report says that “[f]or a program to be considered ‘adequate’ it must have on-street training.” 3 Adequacy Report at B-5, J.A. 216. The unsurprising assumption underlying this statement is that the best way to enhance safety among truck drivers is to ensure practical but supervised experience handling heavy vehicles. FMCSA’s final rule flouts this premise. Quite clearly, the four topics it embraces do not touch on the operational skills of driving a heavy truck. Nothing in the final rule, the administrative record, or even the arguments presented in litigation, suggests any reason to believe that the agency changed course on the basis of evidence that the Adequacy Report’s conclusions were faulty. The fi*345nal rule’s purported responsiveness to the Adequacy Report is therefore flatly contradicted.
From a purely economic perspective, the agency’s disregard of the Adequacy Report is baffling in light of the evidence in the record. Instituting a training regimen along the lines sketched in the Model Curriculum would, according to the agency’s own calculations, produce benefits far in excess of costs. As noted earlier, the program’s estimated 10-year cost of between $4.19 billion to $4.51 billion would yield a benefit ranging from $5.4 billion to $15.27 billion, depending on analytic assumptions. See J.A. 258-62. The cost-benefit analysis in favor of the final rule, however, lends no support to FMCSA’s position. In the final rule, FMCSA says practically nothing about the projected benefits. After running through the costs of mandating its program, the agency suggests that, to be cost-beneficial, the rule would need to prevent 201 crashes by the 32,400 entry-level drivers affected by its provisions each year, representing approximately a five percent reduction in crash rates. 69 Fed. Reg. at 29,400. But the discussion cites no evidence that the final rule would achieve that goal.
FMCSA’s Final Regulatory Evaluation, which was issued to explain the new rule, underscores the frailty of FMCSA’s analysis. In asserting that the new rule will generate a sufficient benefit, FMCSA says: “A 4.7 percent reduction in crashes for this group would appear plausible when measured against the estimates of potential crash reduction measured in studies cited in the [Notice of Proposed Rulemaking], the [Advanced Notice of Proposed Rule-making], and its accompanying regulatory evaluation.” Regulatory Evaluation, Final Rule, Minimum Training Requirements for Entry-Level Commercial Motor Vehicle Operators at 21, reprinted, in J.A. 444. This makes no sense, because the studies to which this statement alludes are those that measured the effect of more substantial training.
Thus, from the premise that a particular method of driver training reduces crashes, the agency infers that anything it calls “driver training” will reduce crashes. This is patently illogical. It is also in direct tension with a specific finding of the Adequacy Report. To explain the correlation found in some studies between training and increased accident rates, the Adequacy Report noted that “researchers have attributed this tendency to the high variability in training quality, indicating that poor training may give the new driver a false sense of confidence in his/her abilities.” I Adequacy Report at 10, J.A. 181. The agency is wrong to assume that its unstudied training program can piggyback on the demonstrated effectiveness of practical, on-the-road training, and its blithe assurance that any training is beneficial ignores the documented risks flowing from subpar training programs.
FMCSA’s efforts to portray the final rule as consistent with the Adequacy Report are fruitless. For example, at oral argument, agency counsel suggested that the CDL program, when coupled with the training requirements of the final rule, will assure adequate driver training. Transcript of Proceedings (“Tr.”) at 20, Advocates for Highway & Auto Safety v. Fed. Motor Carrier Safety Admin., No. 04-1233 (D.C.Cir.Sept. 12, 2005). In particular, counsel pointed to a section of the Adequacy Report that states: “One possible outcome ... could be a hybrid program, i.e., a combination of the Training— and Performance-based approaches that embodies the advantages of each.” See 1 Adequacy Report at 13, J.A. 184. But FMCSA’s invocation of that language misrepresents the discussion of “hybrid programs” presented in the record. First, as already *346noted, the term “training” in the Adequacy Report did not refer to instruction in any set of subject areas imaginable. It had specific content — namely, the kind of training outlined in the Model Curriculum. And “performance-based approach” certainly did not mean the existing CDL program. Rather, under a performance-based approach, “drivers would be required to pass more comprehensive knowledge and skill tests than are presently required to obtain a Commercial Drivers License (CDL).” Id. at 12, J.A. 183. The deficiencies of counsel’s attempt to characterize the agency rule as a “hybrid program” are readily apparent. Whatever role the CDL plays, it is irrelevant to fulfilling the agency’s mandate under IS-TEA. The CDL requirements were well established by the time the Adequacy Report concluded, unequivocally, that the private sector was not effectively ensuring adequate training. The notion that the final rule “responds” to the Adequacy Report because it works in tandem with existing CDL standards is unpersuasive.
FMCSA’s main strategy in defending the final rule is to suggest that it is the first installment of an incremental program that will fulfill its statutory obligations. This is entirely unconvincing. Agencies surely may, in appropriate circumstances, address problems incrementally. See Mobil Oil Exploration & Producing S.E., Inc. v. United Distribution Cos., 498 U.S. 211, 230-31, 111 S.Ct. 615, 112 L.Ed.2d 636 (1991). However, in this case, FMCSA has not shown that its action inaugurates a program designed to tackle the concerns of the Adequacy Report. Rather, the final rule points to some irrelevant initiatives which FMCSA self-servingly characterizes as part of the agency’s “overall ... effort to improve its driver safety programs.” See Minimum Training Requirements, 69 Fed.Reg. at 29,385. None of these programs involves CMV training. The cited initiatives include: (1) considering whether to modify the CDL test and whether the test can be administered more cost-effectively; (2) identifying the costs and benefits of a graduated license system, an inquiry mandated by another section of ISTEA; (3) publishing an interim final rule intended to heighten awareness of safety regulations among motor carriers; and (4) administering a grant program, in place since 1984, that provides financial assistance to states in aid of “roadside inspections and other enforcement activities designed to improve CMV safety.” Id. at 29,385-86. These so-called “initiatives” do not reflect concrete regulatory proposals to address the training problems identified in the Adequacy Report. Indeed, the grant program predates ISTEA and the Adequacy Report. Even on their own terms, two of the initiatives are so speculative that they may come to nothing. The agency has thus presented no reason to believe that these initiatives, in concert with the new rule at issue here, will address the deficiencies identified by the Adequacy Report.
In short, the record in this case shows that the agency entirely failed to consider important aspects of the CMV training problems before it; it lai’gely ignored the evidence in the Adequacy Report and abandoned the recommendations of the Model Curriculum without reasonable explanation; and it adopted a final rule whose terms have almost nothing to do with an “adequate” CMV training program. FMCSA simply disregarded the volumes of evidence that extensive, on-street training enhances CMV safety. FMCSA’s action was thus arbitrary and capricious under § 706(2)(A).
C. UMA’s Arguments
Petitioner UMA, in addition to endorsing the broad criticisms of the final rule, argues that the agency acted arbi*347trarily and capriciously in issuing a rule that applied uniformly to the trucking and motor coach industries. According to UMA, evidence that the motor coach industry has a superior safety record, coupled with the motor coach industry’s economic vulnerability, makes uniform treatment unfair and irrational.
1. Waiver
FMCSA maintains that three of UMA’s arguments — (1) that the industry should be given an opportunity to work with the agency to develop a better training program, (2) that insurance carriers might require training of all drivers rather than just entry-level drivers, and (3) that motor coach operators will be exposed to liability — are waived because they were not raised before the agency prior to the promulgation of the final rale. UMA responds that, in light of Sims v. Apfel, 530 U.S. 103, 120 S.Ct. 2080, 147 L.Ed.2d 80 (2000), there can be no “waiver.”
The agency correctly asserts that, as a general proposition, the applicable case law emphasizes the need for parties seeking judicial review of agency action to raise their issues before the agency during the administrative process in order to preserve those issues for judicial review. See, e.g., United States v. L.A. Tucker Track Lines, Inc., 344 U.S. 33, 37, 73 S.Ct. 67, 97 L.Ed. 54 (1952) (“Simple fairness to those who are engaged in the tasks of administration, and to litigants, requires as a general rule that courts should not topple over administrative decisions unless the administrative body not only has erred but has erred against objection made at the time appropriate under its practice.”). However, Sims indicates that this administrative-waiver doctrine does not represent an ironclad rule. And, as a general matter, a party’s presentation of issues during a rulemaking proceeding is not a jiirisdictional prerequisite to judicial review. See Avocados Plus Inc. v. Veneman, 370 F.3d 1243, 1248 (D.C.Cir.2004) (Courts “presume exhaustion is non-jurisdietional unless ‘Congress states in clear, unequivocal terms that the judiciary is barred from hearing an action until the administrative agency has come to a decision.’ ”) (quoting I.A.M. Nat’l Pension Fund Benefit Plan C v. Stockton Tri Indus., 727 F.2d 1204, 1208 (D.C.Cir.1984)) (emphasis added).
UMA relies on Sims to argue that it is inappropriate to apply the general principles of issue waiver to administrative rule-making. Sims involved a Social Security claimant who was denied disability benefits and then requested that the Social Security Appeals Council (“Council”) review her claims. Sims, 530 U.S. at 105, 120 S.Ct. 2080. After the Council denied review, Sims filed suit in federal district court. She lost and, on appeal, the Fifth Circuit held that two of the three arguments she pressed were not reviewable on the merits, because she had not raised them before the Council. Id. at 106, 120 S.Ct. 2080. The Supreme Court reversed. The Court began its analysis by noting “that requirements of administrative issue exhaustion are largely creatures of statute,” and that most of the Court’s cases refusing to consider arguments initiated in litigation involved specific statutory directives. Id. at 107-08, 120 S.Ct. 2080. The Court noted that when neither statute nor regulation required issue exhaustion, the Court has occasionally imposed its own exhaustion hurdle. But in those cases, the administrative context was critical: “The basis for a judicially imposed issue-exhaustion requirement is an analogy to the rule that appellate courts will not consider arguments not raised before trial courts.” Id. at 108-09, 120 S.Ct. 2080. The application of issue exhaustion “depends on the degree to which the analogy to normal adversarial litigation applies in a particular administrative proceeding.” Id. at 109, 120 S.Ct. *3482080. Where the parties must present and develop issues, the adjudicative model is apt and issue exhaustion is appropriate. “Where, by contrast, an administrative proceeding is not adversarial, ... the reasons for a court to require exhaustion are much weaker.” Id. at 110, 120 S.Ct. -2080. A four-justice plurality in Sims pointed out that “Social Security proceedings are inquisitorial rather than adversarial,” that they are highly informal, and that many claimants are not represented by attorneys, thus making issue exhaustion inapposite. Id. at 110-12, 120 S.Ct. 2080.
UMA contends that, in light of the Court’s holding in Sims, FMCSA’s waiver argument is misplaced. Pet’r Reply Br. at 21 (“Rulemakings are classic examples of non-adversarial administrative proceedings.”). This argument is not unreasonable, because there appears to be no statute or regulation compelling exhaustion in advance of judicial review, and no argument has been made analogizing the agency’s rulemaking to adjudication.
The difficulty that UMA faces, however, is that the case law post-dating Sims gives little support to its position. For example, in Appalachian Power Co. v. EPA, 251 F.3d 1026 (D.C.Cir.2001), the court considered a host of challenges to an Environmental Protection Agency (“EPA”) emissions regulation. The court noted: “It is black-letter administrative law that ‘[a]b-sent special circumstances, a party must initially present its comments to the agency during the rulemaking in order for the court to consider the issue.’ ” Id. at 1036 (quoting Tex Tin Corp. v. EPA, 935 F.2d 1321, 1323 (D.C.Cir.1991)) (alterations in Appalachian Power). Similarly, in National Wildlife Federation v. EPA, 286 F.3d 554 (D.C.Cir.2002) (per curiam), the court upheld a final rule regulating the bleaching process used by paper mills. Among the many challenges brought by the petitioner and rejected by the court was one that “neither [petitioner] nor any other party before the agency raised ... during the administrative phase of the rulemaking process.” Id. at 562. Citing the “well established” principle that “issues not raised before the agency are waived and this Court will not consider them,” the court refused to transgress the “near absolute bar against raising new issues — factual or legal — on appeal in the administrative context.” Id.
Neither Appalachian Power nor National Wildlife Federation mentions Sims or seeks to determine whether the rule-making proceedings in those cases were analogous to adversarial litigation. Sims was addressed, however, in National Mining Ass’n v. Department of Labor, 292 F.3d 849 (D.C.Cir.2002). There the court declined to consider a challenge to the Department of Labor’s regulations under the Black Lung Benefit Act, because the petitioner “failed to raise it during the notice-and-comment period.” Id. at 874 (citing Nat’l Wildlife Fed’n, 286 F.3d at 562). The court found Sims “inapplicable, for it addresses issue exhaustion, not issue waiver.” Id.
The distinction between “issue exhaustion” and “issue waiver” is illusive, to say the least. Indeed, both terms appear in the case law without apparent distinction, and they are sometimes treated as if synonymous. Compare Nuclear Energy Inst., Inc. v. EPA, 373 F.3d 1251, 1297 (D.C.Cir. 2004) (per curiam) (“It is a hard and fast rule of administrative law, rooted in simple fairness, that issues not raised before an agency are waived and will not be considered by a court on review.”) (citing L.A. Tucker Truck Lines, 344 U.S. at 37, 73 S.Ct. 67), ivith Petroleum Commc’ns, Inc. v. FCC, 22 F.3d 1164, 1169 (D.C.Cir.1994) (discussing “the judicially-created requirement of exhaustion, which holds that ‘courts should not topple over administra*349tive decisions unless the administrative body not only has erred but has erred against objection made at the time appropriate under its practice.’ ”) (quoting L.A. Tucker Truck Lines, 344 U.S. at 37, 73 S.Ct. 67). At least one of our sister circuits has explicitly recognized the interchangeability of the two terms. See Coal, for Gov’t Procurement v. Fed. Prison Indus., Inc., 365 F.3d 435, 461-62 (6th Cir. 2004) (“The administrative waiver doctrine, commonly referred to as issue exhaustion, provides that it is inappropriate for courts reviewing agency decisions to consider arguments not raised before the administrative agency involved.”). Indeed, in Sims itself, the Court stated the question presented in terms of waiver. Sims, 530 U.S. at 105, 120 S.Ct. 2080 (“The question is whether a claimant pursuing judicial review has waived any issues that he did not include in that request.”) (emphasis added); see also id. at 115, 120 S.Ct. 2080 (“No one claims that any established exception to this ordinary ‘exhaustion’ or ‘waiver’ rule applies.”) (Breyer, J., dissenting).
The obvious point of the court’s judgment in National Mining Ass’n is that a party will normally forfeit an opportunity to challenge an agency rulemaking on a ground that was not first presented to the agency for its initial consideration. There are two reasons for this. First, the courts are not authorized to second-guess agency rulemaking decisions; rather, the role of the court is to determine whether the agency’s decision is arbitrary and capricious for want of reasoned decisionmaking. See State Farm, 463 U.S. at 43, 103 S.Ct. 2856 (“The scope of review under the ‘arbitrary and capricious’ standard is narrow and a court is not to substitute its judgment for that of the agency.”). Therefore, it is unsurprising that parties rarely are allowed to seek “review” of a substantive claim that has never even been presented to the agency for its consideration. Second, as noted above, “[sjimple fairness ... requires as a general rule that courts should not topple over administrative decisions unless the administrative body ... has erred against objection made at the time appropriate under its practice.” L.A. Tucker Truck Lines, 344 U.S. at 37, 73 S.Ct. 67.
The bottom line here is that, no matter how we characterize the result, UMA forfeited the opportunity to seek judicial review of its claims that the industry should be given an opportunity to work with the agency to develop a better training program, that insurance carriers might require training of all drivers rather than just entry-level drivers, and that motor coach operators will be exposed to liability. Because UMA did not raise these contentions during the rulemaking, and because they are not the kind of clear points that an agency must consider sua sponte, see State Farm, 463 U.S. at 51, 103 S.Ct. 2856 (a rulemaking cannot be found wanting simply because an agency fails to address every alternative “ ‘thought conceivable by the mind of man’ ”) (quoting Vermont Yankee Nuclear Power Corp. v. Natural Res. Defense Council, Inc., 435 U.S. 519, 551, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978)), the FMCSA did not act arbitrarily or capriciously in failing to consider these claims.
2. UMA’s Remaining Claims
Four arguments that UMA raised during the rulemaking process and pursued on appeal remain before this court. However, none of these arguments have merit. First, UMA insists that its superior safety record warranted exempting the industry from the final rule. Second, petitioner accuses the agency of arbitrarily ignoring UMA’s pleas to integrate the four areas of training identified by the final rule into the CDL curriculum. Third, UMA contends that because the industry *350lacked specialized training schools, it would suffer a disproportionate cost in complying with the final rule. Fourth, UMA objected to the agency aggregating the estimated crash costs derived from all of the regulated industries. Since the motor coach industry has a superior accident record and contributes less to overall accident-related costs, UMA posits, it was methodologically unsound to calculate the rule’s benefits without disaggregating the costs associated with each regulated sector.
None of these criticisms demonstrate a failure of decisionmaking sufficient to discard the rule as arbitrary and capricious. Because they are so frail, we devote only brief attention to their specific shortcomings. First, it was not unreasonable for FMCSA to decline the UMA’s request for exemption. Though the Adequacy Report did find that the motor coach industry had a better safety record than the heavy truck industry, it still found the industry’s training record to be inadequate. See Minimum Training Requirements, 69 Fed.Reg. at 29,389. Next, rejecting UMA’s suggestion that the training topics be integrated into the CDL requirement was a permissible policy choice. The agency was free to determine that employers, rather than state administrators, should bear the cost of the final rule. See id. at 29,388. Third, UMA’s argument that it cannot afford to implement the final rule because the industry lacks a training infrastructure makes little sense, considering that UMA endorses the broader proposition that adherence to the Adequacy Report will appropriately involve more extensive training requirements than are now imposed by the final rule. Finally, UMA’s objection to the agency’s aggregation of crash costs seems confused. The agency used the average cost of crashes involving large trucks to estimate the average cost of all crashes affected by its rule. Id. at 29,397. Citing the motor coach industry’s lower frequency of accidents, UMA seeks to discredit the agency’s focus on the costs per accident generated by the trucking industry. Because UMA has not shown that motor coach accidents are less costly when they occur, the association has not demonstrated a methodological problem with the cost-benefit analysis beyond the ones already identified in this opinion.
D. The Appropriate Remedy
Petitioners’ opening brief oscillates with respect to the remedy sought. Compare Pet’r Br. at 46 (requesting vacatur) with id. at 60 (requesting remand). We are convinced that the final rule should be remanded, but that it should remain in effect while the agency crafts an adequate regulation. While unsupported agency action normally warrants vacatur, Ill. Pub. Telecomm. Ass’n v. FCC, 123 F.3d 693, 693 (D.C.Cir.1997), this court is not without discretion. “The decision whether to vacate depends on the seriousness of the order’s deficiency ... and the disruptive consequences of an interim change that may itself be changed.” Allied-Signal, Inc. v. U.S. Nuclear Regulatory Comm’n, 988 F.2d 146, 150-51 (D.C.Cir.1993) (internal quotations omitted).
Advocates conceded at oral argument that the agency’s rule will do no affirmative harm, arguing only that it does not go far enough. See Tr. at 12-13. Accordingly, they raise no objection to our leaving the current rule in place, requesting only that we require the agency to engage in further rulemaking. UMA maintained that allowing the final rule to remain in effect would harm the motor coach industry, but its argument is unconvincing. Stressing the industry’s supposed economic vulnerability, UMA claimed that it cannot afford to implement the final rule. Id. at 14. But as noted earlier, this argument is hard to credit when UMA has been litigating in favor of a far more extensive — and presumably costlier — training *351regimen. UMA recognizes this tension, and insists that if the industry must bear the cost, it needs “a return on investment” that only serious training can provide. Id. Like Advocates, UMA advances no argument that this rule will have a detrimental effect on safety. That leaves only its self-contradictory argument about costs, which is not enough to convince us to vacate a rale that, while plainly inadequate, may do some good, if it does anything at all.
m. Conclusion
The petitions for review are granted as. indicated above, and the case is remanded to the agency for further consideration consistent with this opinion.