concurring:
While I agree with the majority’s treatment in Parts I, II, III, and V dealing with the standard of review, the applicability of § 18(e)(2), and the bias challenge, I am writing separately to express concern about the majority’s wholehearted embrace in Part IV of “the Commission’s assessment of the costs and benefits of the [known defects] proposal [as] a reasonable one, supported by the requisite substantial evidence.” Maj. op. at 426.
The critical finding in the Commission’s explanation of its decision to reverse its earlier position and to eliminate the known defects disclosure provision from the Used Car Rule is that “dealers do not ordinarily possess knowledge about specific defects” in the used cars they sell. 49 Fed.Reg. at 45,712. This finding is central to the Commission’s ultimate conclusion that the potential benefits of the provision are so minimal that they do not outweigh its costs. Yet the Commission’s rationale for promulgating the weaker rule it finally did adopt rests on a seemingly contradictory premise that substantial numbers of dealers are presently misrepresenting the condition of used cars to customers. See id. at 45,702 (“many used car dealers have knowingly misrepresented the mechanical condition of the cars they sell and thereby cause substantial injury to consumers”). Indeed such a premise is a necessary predicate of a rule that requires stickers to warn potential buyers not to depend on oral assurances of dealers, to check systems for themselves, and to consider obtaining independent inspections. Were dealers content to sit back and let customers judge used cars for themselves, the rule would make no sense.
In order to adequately explain its rejection of the known defects disclosure requirement, the Commission had to distinguish between the pervasiveness of dealer misrepresentations based only on general knowledge of a car’s condition and those based on knowledge that specific defects exist. There is admittedly conflicting evidence in the record concerning the thoroughness with which dealers inspect used cars and the extent to which those inspections reveal specific defects. In the end, however, I cannot say that the Commission could not properly conclude on the rule-*214making record that dealers “ordinarily” lack knowledge about specific defects. I do feel compelled, however, to point out that there was considerable evidence that in a great many cases dealers do know about specific defects in the cars they sell. The Commission found in 1981 that “[dealers acquire a great deal of information regarding the mechanical condition of vehicles they sell,” 46 Fed.Reg. at 41,354, and it listed a number of sources through which dealers acquire that information:
The record shows that dealers go to great lengths to learn the mechanical condition of the cars they sell. They obtain knowledge of defects from various sources. The record reflects that, at auctions, dealers may inspect after purchase and in fact, may rescind or renegotiate if problems found are severe. After purchase by dealers, additional defects are discovered during further inspections, appearance reconditioning and repairs. In addition, dealers also become aware of defects through third-party service contract company inspectors, ... through sellers, and through selling or servicing the car previously.
Id. at 41,342. It is, of course, next to impossible to quantify, on a record devoid of any statistics or sampling about the degree of dealer knowledge, whether more dealers do or do not know about specific defects.1 On the anecdotal evidence that did exist, I must give the Commission the benefit of the doubt, but not without reservations.
The majority and the Commission attempt to bolster sparse evidence on dealer knowledge with the theory that requiring disclosure of known defects would create a disincentive for dealers to inspect cars for specific defects, so that “the degree of information dealers would have concerning specific defects under the disclosure provision would be no more, though it might well be less, than what they currently possess.” Maj. op. at 423. Of course, this disincentive, if it exists, cannot be counted as a cost of the known defects disclosure proposal, since without the requirement consumers will get no benefit from any information dealers may have, anyway. But once we accept the finding that “ordinarily” dealers do not know of specific defects, it becomes hard to label as arbitrary or capricious the Commission’s failure to require disclosure from dealers that do know about defects unless it appears that the benefits to consumers would be substantial enough to make worthwhile a *215rule for the extraordinary situation. This record simply does not provide that kind of evidence.
Nevertheless, other parts of the Commission’s rationale, and particularly its discussion of the potential costs of the known defects disclosure provision, strike me as so far-fetched that I cannot acquiesce silently in the majority’s acceptance of them.
The Commission concluded, without citing any record evidence, that “the defect disclosure provision included in the 1981 Rule may confuse consumers and cause them to make inaccurate assumptions about the condition of a car after reading the defect disclosure.” 49 Fed.Reg. at 45,715. Although it acknowledged that the window sticker mandated by the 1981 Rule “contained a warning that the absence of a disclosed defect does not necessarily mean that the car is free from defects,” the Commission stated that “there is no evidence that this warning would be effective.” Id. at 45,716 n. 301. In fact, there was evidence of the effectiveness of the 1981 warning in preventing consumer confusion about the meaning of the defects disclosure statement in the record. Market research conducted by the Commission before the adoption of the 1981 Rule compared consumer reaction to two versions of the defects disclosure section of the window sticker. One of the versions included the sentence “If nothing is listed, the car is not necessarily free of defects,” which was incorporated into the window sticker required by the 1981 Rule. The study found that this sentence, along with the reference to the list of defects that must be disclosed if known (which was also included in the window sticker) “left no doubt in the consumers’ minds” that the listing of defects was not necessarily complete. Exploratory Research Into Consumer Attitudes Toward the Used Car Buyers Guide 17-18 (1981). Moreover, the Commission itself apparently relied on this evidence in 1981, when it noted that “the disclosure is drafted so that consumers will be aware of the fact that dealers have no obligation to inspect prior to purchase.” 46 Fed.Reg. at 41,343. The Commission offered no explanation in 1984 of why it ignored this market research evidence in order to reach the opposite conclusion that consumers would be confused by the defects disclosure.
The Commission also argued that the defects disclosure requirement would reduce the effectiveness of other sections of the rule by “encouraging [consumers] to focus their attention on dealer-controlled information about a car’s mechanical condition” rather than seeking independent inspections. 49 Fed.Reg. at 45,716. The Commission cited no evidence for this conclusion and, in fact, the record suggests that a purchaser who receives a defects disclosure statement may be more likely to seek an independent inspection. See Baseline Survey at 26, Table 10 (19% of Wisconsin used car purchasers had cars inspected before purchase, as opposed to 17% in other states). But even if the Commission is correct in assuming that the defects disclosure requirement will discourage consumers from seeking third-party inspections, this fact represents a significant cost only if consumers would be likely to have used cars inspected before purchase in the absence of the requirement. The Commission acknowledged that currently “few consumers actually seek independent inspections by a qualified mechanic,” 49 Fed.Reg. at 45,706, and it presented no evidence that consumers would be more likely to obtain such inspections after adoption of the rule. In fact, the record shows that the principal reasons why few customers currently obtain pre-purchase inspections are that dealers generally refuse to allow independent inspections and that, even when inspections are allowed, consumers are not willing to accept the inconvenience and expense that they entail. See 49 Fed.Reg. at 45,701 n. 157 (“the record demonstrates that off-premises inspections are not allowed on a regular basis”); Final Staff Report at 7 (“pre-sale inspection by an independent mechanic is inconvenient and inefficient to the consumer and unpalatable to even the honest dealer, who must either risk the theft of his vehicle or incur the cost of an employee’s time to safeguard it”); id. at 87-89 (“considerable information holds that off-premises inspections are not allowed”). None of these factors that currently make *216prepurchase inspections unlikely is affected by the 1984 Used Car Rule. Thus, even after adoption of the rule, most consumers will presumably continue not to obtain independent inspections. What, if anything, the new rule will do for consumers is obscure at best, and this part of the Commission’s rationale for dropping the known defects proposal insubstantial.
I am well aware that the Commission’s decision to eliminate the defects disclosure provision from the Used Car Rule is rooted in a fundamental shift in the agency’s view of how the public interest may best be served. Consumer self-help and a self-correcting free market have replaced aggressive regulation as the order of the day. Nonetheless, the Supreme Court has reaffirmed only recently that “ ‘[a]n agency’s view of what is in the public interest may change, either with or without a change in circumstances. But an agency changing its course must supply a reasoned analysis....’” Motor Vehicle Manufacturers Association v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 57, 103 S.Ct. 2856, 2874, 77 L.Ed.2d 443 (1983) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C.Cir.1970)). This one makes it, but just under the wire.
. Neither of the two empirical studies cited by the "Wisconsin Study” and the "Baseline provided any direct evidence on the question of dealer knowledge of specific defects. The Commission did rely on the studies to support its conclusion that the defects disclosure requirement would provide negligible benefits to used car buyers. Using data from the two studies, the Commission compared used car buyers in Wisconsin after the enactment of a statute requiring defects disclosure with buyers in other states and also preenactment and postenactment buyers within Wisconsin. The Commission concluded that "the expected beneficial effects of a defect disclosure requirement were not achieved in Wisconsin.” 49 Fed.Reg. at 45,714. There are significant differences, however, between the Wisconsin statute and the defects disclosure proposal that was before the Commission. The Wisconsin law required dealers to present a defects disclosure statement to consumers before the sales contract was signed but did not require that defects be listed on a window sticker, as the rule adopted by the Commission in 1981 did. Id. at 45,714 n. 292. The Wisconsin Study reported that "after the disclosure law went into effect, only 16.7 percent of Wisconsin's consumers who purchased a used motor vehicle from a dealer had access to the disclosure statement before deciding to purchase the vehicle.” Wisconsin Study at 37. The authors of the Study noted that "the timing and manner of disclosure may be as important as disclosure itself," and concluded that "[t]here is little question that seeing the disclosure statement prior to making the decision to purchase a vehicle is in the consumer’s best interests.” Id. at 37, 39. Because the timing of disclosure can significantly affect its benefits for consumers, the validity of using studies of the effects of the Wisconsin statute to evaluate the potential benefits of the defects disclosure proposal that was before the Commission is doubtful. Nevertheless, as the Commission points out, see 49 Fed.Reg. at 45,714 n. 292, the Wisconsin Study and the Baseline Survey provided the only empirical evidence in the record concerning the effects of a defects disclosure requirement, and it is difficult to fault the Commission for considering this evidence.