Petitioner D. D. Bean & Sons Co. (Bean) is a manufacturer of paper bookmatches with its principal place of business in Jaffrey, New Hampshire. Pursuant to 15 U.S.C. § 2060(a), Bean seeks review of the Consumer Product Safety Commission's matchbook safety regulations promulgated on May 4, 1977, with a scheduled effective date of May 4, 1978. See 42 Fed.Reg. 22656.

The Commission's matchbook safety proceedings were officially commenced in 1974. 39 Fed.Reg. 32050 (Sept. 4, 1974). As authorized by 15 U.S.C. § 2056(d), the Commission accepted the offer of the American Society for Testing and Materials to develop matchbook safety standards. In early 1975, the Society transmitted its proposals to the Commission. In 1976 the Commission published a modified version of the Society's proposal as a proposed Safety Standard. 41 Fed.Reg. 14112 (April 1, 1976). The Commission stated that the proposed regulation was "designed to reduce or eliminate the following unreasonable risks of death or serious injury . . . .

"1. Burn injuries sustained by children and others, including mentally or physically impaired persons, who play with or otherwise improperly use matchbooks.

"2. Burn injuries sustained by persons who use bookmatches that spark or have delayed ignition.

"3. Eye injuries sustained by persons who use bookmatches that fragment and cause particles from such matches to lodge in a person's eye."4. Burn injuries sustained by persons who use bookmatches that, when struck, ignite the remaining matches in the matchbook.

"5. Burn injuries sustained by persons, including children, using bookmatches that after ignition have dropped on exposed parts of the body or that have been dropped on and have ignited clothing because such bookmatches have failed to extinguish in time to avoid such injuries.

"6. Burn injuries sustained by persons from fires that have resulted from unexpected ignition of bookmatches with no deliberate action by the user.

"7. Burn injuries that have been sustained by persons from fires that have been set by the afterglow of extinguished bookmatches."

Id. After receiving comments from the industry and members of the public, the Commission modified the proposed standard. The Commission deleted, as economically and technologically unfeasible, a proposal for a mandatory, child-resistant matchbook design which had been suggested to meet the first hazard enumerated by the Commission.1 The Commission also deleted a proposed requirement that matches self-extinguish within several seconds, a rule developed to meet the fifth hazard. 42 Fed.Reg. at 22660-61 (May 4, 1977). The Commission's final rule contained the following design or "general" requirements:

"(a) The friction shall be located on the outside back cover near the bottom of the matchbook.

"(b) The cover shall remain closed without external force.

"(c) No friction material shall be located on the inside of the cover where possible contact with the matchheads may occur during ordinary use.

"(d) There shall be no bridge(s) or broken bridge(s).

"(e) No matchhead in the matchbook shall be spilt, chipped, cracked, or crumbled.

"(f) No portion of any matchhead shall be outside the matchbook cover when the cover is closed.

"(g) No part of a staple or other assembly device for securing the cover and combs shall be within or touching the friction area.

"(h) A staple used as an assembly device for securing the cover and combs shall be fully clinched so that the ends are flattened or turned into the cover."

Id. at 22669. The rule also listed the following "performance requirements" with which matchbooks must comply:

"A matchbook is defective . . . if it has, when tested in accordance with § 1202.6 (of the Rule), any of the following performance defects:

"(a) Delayed ignition exceeding 2 seconds.

"(b) A splint that separates into two or more pieces.

"(c) A matchhead that produces fragments as defined in § 1202.3(h).

"(d) A matchhead that falls off.

"(e) Friction that burns or produces fragments as defined in § 1202.3(h).

"(f) Afterglow exceeding 5 seconds or the reappearance of a visible flame after the first extinction of the original flame."

Id. The Commission left to later rule-making the establishment of regulations for manufacturers' individually devised testing and certification programs:

"The certification rule, which will subsequently be proposed by the Commission for comment, will establish the parameters of a reasonable testing program. A manufacturer would be required to certify that his product complies with the standard, based on such a program. It may be that a reasonable testing program will include test procedures which are different from those set out in the standard, but the Commission will test for compliance by using the procedures in the standard."Id. at 22664. Between the effective date of the Rule presently under review and the establishment of an approved certification program a manufacturer who sells matches in violation of the general or performance requirements subjects himself to civil or criminal penalties. See 15 U.S.C. §§ 2068-70. The Commission will test for violations in accordance with detailed testing procedures that are set out at length in the Rule. 42 Fed.Reg. at 22669. The testing procedures call for a visual, post-manufacture inspection of matchbooks to insure compliance with the "general" requirements governing such matters as friction location, splitting or bridging of matchheads, etc. "Performance" defects are to be minimized by more elaborate testing. Samples from a lot are to be conditioned in special ovens, then tested under laboratory conditions for the incidence of fragmentation, delayed ignition, etc. Id.

As required by 15 U.S.C. § 2058(c)(1)(C), the Commission also made findings concerning the economic impact of the Rule:

"The Commission estimates that manufacturing cost increases as a direct or indirect effect of this standard will be modest for the industry as a whole. Such increases will tend to be concentrated in one-time costs to complete change-over to reverse friction, and in costs to establish and implement testing programs and certification procedures.

"Because some 80-90 percent of the matchbooks produced annually are given free to consumers, there is not likely to be any direct cost impact on the consumer as a result of the standard. Some proportion of increased manufacturing costs will be passed on to the institutions and business enterprises that purchase matchbooks for promotional purposes. To the extent that increases in advertising and promotional costs may be reflected in higher prices for goods and services sold by these businesses, there may be indirect cost effects on consumers. If so, such impacts would likely be small, if not imperceptible.

"For the 10-20 percent of matchbooks that are purchased at retail by consumers some proportion of any manufacturing cost increases may be passed on to the consumer. A resulting increase in retail prices for such matchbooks will be small, no more than a few cents per box of 50 matchbooks.

"The Commission finds that the standard will not have impacts of significant magnitude on the availability of matchbooks. Although some institutions and business enterprises may reduce their matchbooks purchases or eliminate them in response to any increased price of matchbooks, the large number of such purchasers, and the large volume purchased annually, are such that curtailment of purchases by some businesses is likely to have very small effects on the total number of matchbooks available to U.S. consumers."

42 Fed.Reg. at 22668. The evidence in the record supporting the Commission's findings as to the cost of compliance with the Rule consists of a report of an economic consultant. He estimated that the capital investment required in order to produce matchbooks complying with the Rule would amount to $808,000 and that the industry's annual cost for testing procedures would total $4.3 million.2 For "resale" matchbooks, those sold or given away in retail stores, he estimated a 5% price increase and a 5% decline in sales resulting from the Rule. "Special reproduction" matchbooks would experience a price increase of 2.9% with a scarcely measurable decline in volume sold. Higher costs are predicted by industry representatives. One match manufacturer argued to the Commission that price increases of greater than 10% would be experienced and that volume might decline by as much as 25%. Bean's estimated price increases, as submitted to the Commission, were of a similar magnitude.

The Consumer Product Safety Act, 15 U.S.C. §§ 2051 et seq., provides that "any person adversely affected . . . or any consumer or consumer organization" may seek circuit court review of a consumer product safety rule. Id. § 2060(a). Review is to be in accordance with the Administrative Procedure Act, 5 U.S.C. §§ 701 et seq., and "(t)he consumer product safety rule shall not be affirmed unless the Commission's findings under section 2058(c) . . . are supported by substantial evidence on the record taken as a whole." 15 U.S.C. § 2060(c).

Section 2058(c)(1) requires findings with respect to:

"(A) the degree and nature of the risk of injury the rule is designed to eliminate or reduce;

"(B) the approximate number of consumer products, or types or classes thereof, subject to such rule;

"(C) the need of the public for the consumer products subject to such rule, and the probable effect of such rule upon the utility, cost, or availability of such products to meet such need; and

"(D) any means of achieving the objective of the order while minimizing adverse effects on competition or disruption or dislocation of manufacturing and other commercial practices consistent with the public health and safety."

Section 2058(c)(2) provides that the Commission must also find:

"(A) that the rule (including its effective date) is reasonably necessary to eliminate or reduce an unreasonable risk of injury associated with such product;

"(B) that the promulgation of the rule is in the public interest; . . . "

In connection with the requirement that the rule be "reasonably necessary to eliminate or reduce an unreasonable risk of injury," the House Committee Report had the following to say:

"It should be noted that the Commission's authority to promulgate standards under this bill is limited to instances where the hazard associated with a consumer product presents an unreasonable risk of death, injury, or serious or frequent illness. Your committee has not included a definition of 'unreasonable hazards' within this bill. Protection against unreasonable risks is central to many Federal and State safety statutes and the courts have had broad experience in interpreting the term's meaning and application. It is generally expected that the determination of unreasonable hazard will involve the Commission in balancing the probability that risk will result in harm and the gravity of such harm against the effect on the product's utility, cost and availability to the consumer. An unreasonable hazard is clearly one which can be prevented or reduced without affecting the product's utility, cost, or availability: or one which the effect on the product's utility, cost or availability is outweighed by the need to protect the public from the hazard associated with the product. There should be no implication, however, that in arriving at its determination the Commission would be required to conduct and complete a cost-benefit analysis prior to promulgating standards under this act. Of course, no standard would be expected to impose added costs or inconvenience on the consumer unless there is reasonable assurance that the frequency or severity of injuries or illnesses will be reduced."

H.R.Rep. No. 1153, 92nd Cong., 2nd Sess. 33 (1972).

The Commission's present Rule sets out the "general" and "performance" requirements enumerated above. The requirements may be said to be designed to reducing the following supposed hazards: (1) matchbooks which ignite accidentally (addressed by general requirements a, b, c & f, requiring reverse and no stray friction and covers that remain closed without exposed matches); (2) delayed ignition of matches (performance requirement a); (3) split splints (performance requirement b); (4) afterglow and reappearance of a visible flame (performance requirement f); (5) fragmentation (general requirements d, e, g, & h, prohibiting bridged and chipped matches and exposed staples, and performance requirements c, d & e, banning fragmentation).

In ascertaining the sufficiency of the basis for the Commission's Rule, we shall first inquire whether there is substantial evidence for its findings that each of the listed hazards is in fact such. If it is, we shall next inquire whether the requirements addressed to a particular hazard have been shown to be likely to reduce or eliminate it at a reasonable cost. Only after the existence of a hazard and the likelihood of its reduction at a reasonable cost have been established by the Commission may it be said that the requirements are "reasonably necessary." 15 U.S.C. § 2058(c)(2)(A). See Aqua Slide 'N' Dive Corp. v. Consumer Product Safety Commission, 569 F.2d 831, 838-39 (5th Cir. 1978).

With respect to the first hazard, accidentally ignited matchbooks, we find substantial record support for the Commission's statutorily required finding that an "unreasonable risk of injury" from this hazard exists. Both the Commission's computerized injury surveillance system established at selected emergency rooms across the nation and the Commission staff's "in depth" studies of selected reported accidents, established that a sizable number3 of persons are injured annually by matchbooks which accidentally ignite. Such injuries most commonly occur when a match is struck against a friction located on the front of a matchbook. If the matchbook is left open and the match sprays sparks or fragments, the matchbook may ignite. Injury occurs when the matchbook is dropped onto clothing or adheres to the flesh.

We are also satisfied that the Rule's design or "general" requirements geared to meet this hazard are "reasonably necessary" to reduce the risk from flaming matchbooks. Chief among these is the "reverse friction" requirement, patterned after Canada's and some American manufacturers' policy of locating the friction element on the reverse side of matchbooks. While the record contains no detailed empirical data to support the conclusion that reverse friction is safer, the Commission might rationally so conclude since ignition in the case of an open matchbook is obviously made more difficult when the friction is located on the reverse side. The further design requirements, most notably that the cover remained closed and that no stray friction material be found anywhere in a matchbook, are also logically likely to reduce the risk. As the existence of a hazard was established, and as the curative effects of the various design requirements are patent, we do not think that the Commission had to cite to empirical data in support of its finding that the particular requirements were likely to reduce the risk of injury.

This being so we turn finally still reviewing the Commission's actions to guard against accidental ignition of matchbooks to the question of whether the Commission's cost estimates are supported by substantial evidence and, if so, whether the Commission properly determined in light of those costs that the requirements are "reasonably necessary to eliminate or reduce an unreasonable risk of injury . . . ." 15 U.S.C. § 2058(c)(2)(A). Such a determination entails a balancing of the magnitude of the likely reduction in hazards to consumers against the increased cost to consumers and manufacturers. See H.R.Rep. No. 1153, supra.

With respect to the manufacturers' equipment changes necessary to comply with the reverse friction requirements, the consulting economist estimated a one-time modification costing the industry less than one million dollars. While the record does not reveal the marginal increased cost for visual testing for stray friction and open covers, we have no reason to assume that the added cost to manufacturers will be more than minimal. Bean informs us that visual testing is currently the industry norm. In view of the substantial hazard likely to be reduced by the friction and closed cover requirements and the relatively small cost to the industry, we cannot say that the Commission has erred in finding that these aspects of the Rule are "reasonably necessary" to reduce the hazard from accidental matchbook ignitions.

We have much greater difficulty when we turn to the other hazards and the requirements addressed to them. Performance requirements (a), (b) and (f) are intended to guard against delayed ignition, split splints, and afterglow or reappearance of flame. After examining the record with some care, we have found a dearth of evidence substantiating any degree of risk from these particular conditions. Either the conditions themselves have not been shown to exist to any appreciable extent among currently manufactured matches or else they have not been identified as causing any very significant number of injuries.

With respect to delayed ignition, all of the Commission's data identify only two injuries arguably caused by this hazard. Both occurred in the mid-1960s and are described in the "in-depth" studies, without elaboration, as having been caused by delayed ignition. Otherwise, a Commission staff memorandum simply refers in passing to an unidentified report by a Dr. Hahn which presented the view that delayed ignition was linked to injuries. None of Dr. Hahn's supporting data, if any, is presented.

In a similar vein, we can find no evidence at all in any of the Commission's injury studies that split splints have caused injuries. Nor was there any empirical data, as from a sampling by the Commission of presently manufactured matches, to establish that split splints commonly occur. The closest the record comes to evidence of this purported hazard is a statement by the American Society for Testing and Materials that if split or broken splints occurred, they would be hazardous.

The data supporting a finding of a hazard from afterglow or reappearance of visible flame were also sparse. The inclusion of an afterglow ban was justified by the American Society for Testing and Materials with the statement that "(i)t is generally recognized (and uncontroversial) that a match splint that continues to glow after the flame has been extinguished is a hazard." Whether the presence of afterglow would present a hazard is disputed by Bean. Bean also argues persuasively that there is no substantial evidence that afterglow occurs to any measurable degree. Our review of the record satisfies us that Bean is correct on this score. The Commission appears to have developed no empirical data showing that prolonged afterglow is commonly present in matches.4 And the Commission's various sources of data on injuries establish only that one injury was arguably attributable to afterglow. No information is provided as to how the injury was caused or the degree of its severity. A single injury, so inadequately described in terms of cause and degree, is not substantial evidence of an "unreasonable risk of injury". With respect to possible reappearance of flame, the Society simply commented:

"There was no factual evidence presented to indicate that (reappearance of a visible flame) had caused any injuries and no mention was made of the defect in the risk analysis (of matchbook hazards.) However, common sense seems to indicate that a hazard might exist if a flame reappears after a visible flame had apparently been extinguished."

While "common sense" may so indicate, that is not evidence that the reappearance of a visible flame is a condition that exists with respect to any appreciable number of matches, thus requiring correction. Nor is the occurrence of afterglow, delayed ignition or split splints adequately established. The statutory term "unreasonable risk" presupposes that a real, and not a speculative, risk be found to exist and that the Commission bear the burden of demonstrating the existence of such a risk before proceeding to regulate. See Aqua Slide, supra ; 15 U.S.C. § 2058(c)(2)(A); H.R.Rep. No. 1153, supra. While delayed ignition and the other conditions may well pose a threat of injury to consumers if they exist, the Commission has not established as a threshold matter that the conditions occur to any appreciable extent. Without evidence supporting such a finding, the Commission may not force the industry to adopt the program of testing and quality control needed to certify that its products measure up to "performance" standards designed to guard against such supposed hazards. While the cost of such programs remains uncertain, it is clearly not de minimis. The projection of the Commission's economic consultant of the cost of adopting a program of testing and quality control so that manufacturers can certify to their compliance amounts to several million dollars annually. We therefore conclude that the above performance requirements, being without support in the record now before us, cannot be sustained.

The Commission's performance standards are also fashioned to reduce the occurrence of matchhead fragmentation, (c) and (e), and heads breaking off matches, (d). General requirements (d), (e), (g), & (h) are also addressed to the fragmentation hazard. The data from the Commission's survey of hospital emergency rooms and other studies indicate that roughly one-third of reported match-related injuries consist of fragments lodging in the eye. There are also some reported cases where a flaming fragment ignited a victim's clothing. One study indicated that entire matchheads occasionally fly off upon ignition, lodging in the eye or on clothing. We think therefore that there is sufficient evidence in the record to support the Commission's finding that a substantial hazard exists from matchhead fragmentation.5 We also think that the Commission was entitled to adopt the "general" requirements addressed to the fragmentation hazard. These call essentially for assurance that the matchbooks and matches do not contain obvious defects such as broken splints and matchheads and projecting staples which could lead to particles flying off when a match is struck. Compliance with these requirements may be insured by visual testing, a procedure which is presently the industry norm and will be required, in any event, by the requirements geared to the accidental ignition hazard, discussed supra. And the design requirements themselves merely amount to insisting that the manufacturer produce matchbooks and matches that, on their face, do not contain obvious defects that seem logically to relate to a fragmentation risk. Since the cost of testing to ensure compliance with the general requirements geared to meeting the fragmentation hazard will therefore be slight, and the object of the requirements ensuring a properly functioning product seems only reasonable, we think that these requirements can be said to be "reasonably necessary" to reduce the risk from fragmentation.

However, we are not satisfied that the Commission has carried its burden of showing that the "performance" requirements geared to fragmentation are "reasonably necessary to eliminate or reduce" the risk to any significant degree. Performance requirements, unlike design requirements, cannot be tested for visually. The Commission itself concedes "that the principal cost effects on producers will arise from implementation of testing programs." Hence the effectiveness of these performance requirements in reducing risk is a matter of considerable importance.

There is a complete absence from the record of evidence tending to show the causes of fragmentation. The Commission has evidently proceeded on the assumption that fragmentation results from defective manufacturing procedures and will show up upon testing at the plant. However, manufacturers, in submissions to the Commission, asserted that experience showed that post-manufacture factors such as high humidity, perspiration, or misuse by consumers were principally responsible for fragmentation.6 A commission staff member commented,

"It is difficult to predict to what degree the Standard can reduce matchhead fragmentation. It is true that matchheads which are soft at time of manufacture tend to fragment more than a harder matchhead. Similarly, soft friction can lead to fragments. To the extent that these problems can be controlled in manufacture, the Standard should help. Unfortunately, bookmatches which will not fragment at the time of manufacture may fragment later on in use. Patterns of storage and patterns of use can cause fragmentation. If matches become damp and then dry out, the heads will soften. Some match users strike harder than others and the harder they strike, the more likely the head will fragment.

"From the available data, the number of matches lit, per eye injury reported, is on the order of 73 million. Assuming that something like one fragment in a thousand may hit the eye, then on the average, one out of 7,300 matches will fragment. Assuming, conservatively, that all of these fragmenting matches are the result of the method of manufacture, then we can conclude (a) fragmenting matches are fairly rare and (b) it will be virtually impossible by any sort of certification program to force the reduction or even verify this present rate of fragmenting matches. Of what value then is the certification program? The program's major value is in preventing gross manufacturing errors by requiring periodic inspections and performance tests. Thus, if the matchhead composition is mixed improperly such that a whole series of matches might be made which would tend to fragment, the inspection procedures would most likely catch this batch."

In view of the lack of evidence establishing the causes of fragmentation, and of the manufacturers' submissions claiming that post-manufacture handling causes most fragmentation, it is a large assumption that fragmentation is almost wholly attributable to the manufacturing process. Even then, the above-quoted evaluation the only study in the record tending to support the Commission's position on fragmentation establishes only that, at most "gross manufacturing errors" will be detected. But there is no evidence in the record from which the Commission could have determined that "gross" errors, much less occasional defects, occur with any frequency. It is just as possible to hypothesize that post-manufacture handling is the principal cause of fragmentation. We therefore think it speculative for the Commission to find, on the present record, that the proposed fragmentation performance requirements will result in a significant decrease of the fragmentation risk. See Aqua Slides, supra at 843-844. In view of this, the substantial, added cost of testing to insure compliance with the performance requirements cannot be justified.

It may be argued that our conclusion in this regard necessarily entails an evaluation of the likely effectiveness of manufacturers' testing procedures, judicial scrutiny of which would be best postponed until the Commission's testing certification procedures are promulgated. We do not agree. The Commission's criteria with respect to the performance of manufactured matches are binding on the industry upon the effective date of the present Rule. The Commission will test for defects on its own, and a manufacturer will have to take such precautions, including establishing its own testing program, as may be needed to avoid the civil and criminal penalties which a violation entails. From the Commission's own cost estimates, manufacturers may be obliged to devote several million dollars or more to assure continuing compliance with fragmentation and other performance standards. Since the burden on the industry is immediate and the likelihood or unlikelihood of risk reduction also presently measurable, we think that judicial review of the fragmentation rule should not be postponed. Indeed, the Act requires the Commission to make its findings and strike the regulatory balance "prior to promulgating a consumer safety rule". 15 U.S.C. § 2058(c)(1). Moreover, § 2060(c) provides that "(t)he consumer product safety rule shall not be affirmed (by the court) unless the Commission's findings under § 2058(c) are supported by substantial evidence." The Act clearly contemplates that the correctness of the balance struck by the Commission and the adequacy of its findings be judged as of the time of the Rule's promulgation. In the present case we are unable to say that there is substantial evidence to support a reasonable prediction that the fragmentation performance requirements will likely reduce the risk from this hazard. The Commission therefore has not shown that the Rule is "reasonably necessary to eliminate or reduce" the risk from fragmentation. § 2058(c)(2) (A). See Aqua Slide, supra at 837.

To summarize, all of the "general" requirements of the Rule are affirmed. The "performance" requirements are set aside.7

So ordered.