Full opinion text
JERRY E. SMITH, Circuit Judge: In December 1987, the Occupational Safety and Health Administration (OSHA), exercising its authority and responsibility under section 6 of the Occupational Safety and Health Act (“OSH Act” or the “Act”), 29 U.S.C. § 655, promulgated a new standard designed, in relevant part, to eliminate or reduce the substantial risk of fires and explosions in the nation’s grain-handling facilities. In these consolidated petitions, parties representing both labor-union and industry interests challenge the new standard, particularly the scope of one of its provisions that requires grain elevator operators to initiate clean-up procedures whenever grain dust accumulations reach a depth of Vs inch in certain “priority areas” of the facility. The unions contend that the Vs-inch “action level” provision does not go far enough to protect workers: It should extend beyond priority areas to encompass the entire facility, and it should be mandated not only for grain elevator operators but for grain mills as well. Intervenors representing grain mill operators have countered with briefs supporting the Secretary’s decision not to subject grain mills to the action level. The grain elevator operators — or industry petitioners — argue that the action level is unwarranted because its benefits cannot justify its costs, because it cannot be shown to reduce a substantial risk, and because its costs would cause the industry widespread economic dislocation. Finding that the record does not contain substantial evidence to support either (i) OSHA’s economic-feasibility determination or (ii) its decision not to mandate a facility-wide action level for grain elevators, we remand for further consideration on these points. In all other respects, we deny the petitions challenging OSHA’s grain-facilities standard. I. Background. Fires and explosions in grain-handling facilities have been documented for almost two centuries, although such occurrences probably date back to even earlier times, when structures for storing grain in large quantities were first developed. Only recently, however, has come the impetus for systematic investigation and prevention of these disasters. During an eight-day period beginning on December 21, 1977, and ending on December 28, a series of devastating explosions left 59 dead persons and 48 injured. 52 Fed.Reg. 49,592 (1987). Moved by the 1977 catastrophes, numerous government agencies, including OSHA, the National Institute for Occupational Safety and Health (NIOSH), and the Department of Agriculture (USDA) responded to the hazards of grain storage. OSHA, in particular, took several steps. First, it issued a Grain Elevator Industry Hazard Alert and provided free on-site consultation services. 49 Fed.Reg. 997 (1984). It also attempted, albeit with only limited success, to enforce the general-duty clause in section 5 of the OSH Act, 29 U.S.C. § 654(a)(1), and OSHA’s general industry housekeeping regulations, as a means to move the industry toward safer practices. Industry, however, vigorously resisted these enforcement efforts, and employers generally were successful in arguing that OSHA had not proved that the specific condition cited could cause a fire or explosion. Another step taken by OSHA in 1978 was to commission studies of the causes and prevention of grain elevator fires and explosions. Commissions went to the National Academy of Sciences (NAS) and to Charles Kauffman, Ph.D., and Robert Hubbard, former Vice President of Cargill. In addition, USDA and NIOSH contemporaneously conducted their own studies of worker safety in grain elevators and mills. 52 Fed.Reg. 49,592-93 (1987). The reports issued by the NAS, NIOSH, Kauffman and Hubbard, and the USDA were essentially consistent. All recognized that a grain elevator explosion is the result of a concatenation of four essential elements: (1) grain dust must collect in the elevator; (2) the grain dust must be suspended in air inside the elevator at a concentration above the lower explosive limit (L.E.L.); (3) the suspended grain dust must be ignited; and (4) sufficient grain dust to sustain rapid combustion must be in the vicinity of the initially-ignited grain dust. Thus, if the interior of a facility is dusty, a primary or initial explosion that causes only relatively slight damage can shake loose a large, suspended dust cloud. Ignition of this fuel supply by hot dust particles or flame from the primary explosion can then cause a secondary explosion, a series of successively more violent explosions. Pressure waves preceding the flame fronts disperse static dust accumulations into the atmosphere, making even greater amounts of dust available to fuel the flame fronts. This activity has been likened to the expanding ripples that occur on the surface of a still pond after a pebble is tossed into it. Secondary explosions cause the vast majority of injuries and deaths, because they tend to be much more severe than primary explosions. 52 Fed.Reg. 49,592 (1987). Of the major grain elevator explosions occurring between 1979 and 1981, secondary explosions accounted for 96 percent of the property loss, 85 percent of the fatalities, and 91 percent of the injuries. Even relatively small amounts of layered dust can create dust clouds sufficient to cause an explosion, and the more dust that is present, the greater the hazard. The USDA has found, for example, that accumulations as low as Vie inch can initiate a primary explosion, and “[t]he more dust present, the greater the chance of ignition.” NIOSH found that a dust layer Vm inch deep in an enclosure 10 feet in height can create a dust cloud above the LEL when uniformly dispersed. NAS noted that “if accumulations of dust are visible, the elevator has a potential explosion problem. The thicker the layer of dust, the greater are both the possibility of an explosion and the severity of the resulting damage.” Any movement or handling of grain produces grain dust, and this occurs at each point in the grain-handling system from farm to ultimate consumer. In a grain elevator, numerous activities subject the grain to mechanical stress leading to the production of grain dust. Grain is delivered to an elevator by truck, rail, or barge, and is dumped into a pit that feeds a conveyor belt leading to the “boot,” which is called the bottom of the leg. The leg is an enclosed vertical, endless-belt bucket conveyor that elevates the grain and discharges it into the top of a garner, or collection bin, in the elevator “head.” Grain is discharged from the bottom of the garner into a scale bin for weighing. Afterwards, the grain is made to flow out of the bottom of the scale bin onto a belt conveyor that drops the grain, up to 100 feet, into a bin or silo. Grain is later conveyed back into the bottom of the leg, where it may be re-elevated for mixing and redistributed into a silo, or loaded for shipment. Typically, the grain is elevated in the leg and dropped into the silo at least twice. Grain dust does not remain forever airborne but tends to settle on nearly every horizontal surface in the facility, hidden or exposed. Accumulations near the leg are by far the most hazardous, as they contain so many potential ignition sources. As one expert observed, “[i]t is the one piece of equipment with an environment of suspended dust that is continually subject to 'choking' (boot fills with grain and buckets will not turn), electrical faults, bearing failures, mechanical misalignments, ingestion of foreign material, etc.” NAS Report at 46. All of the studies of grain-handling hazards concluded that comprehensive dust control is central to the control of fires and explosions. Kauffman and Hubbard concluded that 11 of the 14 secondary explosions they studied could have been prevented by a good housekeeping program, and that two were attributable to improper installation of the dust-control system. NAS explained that given the impossibility of eliminating all the elements for grain dust explosion, “there is no single, simple process for preventing explosions.” Nevertheless, NAS rated dust control as a first priority for an effective response to the problem: The possibility of dust suspension in concentrations above the lower explosive limit in open areas of an elevator cannot be ignored but can be reduced with relative ease. Layered dust in open work areas and hidden spaces can be reduced to a less hazardous amount through proper housekeeping. Ignition sources cannot be eliminated totally but the probability of their causing an explosion can be reduced. NAS Report at 31. Similarly, the NIOSH report concluded that “[g]ood housekeeping is probably the single most important factor in reducing the risks associated with secondary grain dust explosions....” When possible, dust should be cleaned up whenever visible tracts are recognized. The housekeeping program should address hidden, as well as visible, dust. NIOSH recommended that housekeeping should keep layered dust below Vs inch, even though that is the minimum depth visible to the human eye. In February 1980, OSHA published an Advance Notice of Proposed Rulemaking (“ANPR”), requesting data and comment on present practices in the grain-handling industry, available control technology, and the likely economic impact of a comprehensive grain-handling standard. 45 Fed.Reg. 10,732 (1980). With a view to the span of operations from farmer to consumer, OSHA requested comments on the extent to which a standard should embrace all sectors of the grain-handling industry. In addition, OSHA sought comment on the amount of grain-dust accumulations that should be considered excessive. Id. at 10,-734. Based upon the ANPR comments, the NAS reports, and other materials, OSHA drafted a proposed rule which it circulated to employers, unions, and other interested parties in December 1982. This draft proposal covered all milling operations and all grain elevators, except those on farms. It set out a comprehensive fire and explosion control system closely patterned after NAS’s first-priority recommendations. The system included employee training, routine preventive maintenance, engineering controls, foreign object screens, and a housekeeping program that required dust cleanup whenever accumulations averaged Vs inch over a 200-square-foot floor area anywhere in the facility. After incorporating industry and union comment, OSHA transmitted a draft proposed rule to the Office of Management and Budget (OMB) pursuant to Executive Order No. 12,291, 3 C.F.R. § 127 (1981), reprinted in 5 U.S.C. § 601 note at 431 (1982). In response, OMB questioned whether, in light of recent industry safety practices, there was a need for a standard, and whether OSHA’s draft standard would have the safety benefits OSHA sought. OMB urged OSHA, in any event, to exempt mills and small elevators from any standard and to eliminate the Vs-inch action level housekeeping provision, or at least to offer two alternatives to it: a once-per-shift cleaning, or installation of pneumatic dust collection systems. In January 1984, OSHA published a proposed rule. 49 Fed.Reg. 996 (1984). The housekeeping requirement provided that employers must develop and implement a housekeeping program consisting of a dust control and removal method or combination of methods which will minimize fugitive grain dust accumulations inside grain handling facilities on ledges, floors, equipment, and other exposed surfaces. Id. at 1009. Consistent with OMB’s recommendations, the proposed standard further required that, as part of the housekeeping program, employers do one of three things: (1) undertake dust control when dust reached an action level of Vs inch averaged over a 200-square-foot area; (2) establish a once-per-shift cleaning; or (3) install a pneumatic dust control system. Id. The proposed rule generated 5,000 comments, 11 days of hearings, and numerous submissions of data, as well as post-hearing comments and briefs. 52 Fed.Reg. 49,-594 (1987). A number of participants questioned the need for any standard, arguing that the industry was dealing with the problem effectively on its own, and that the likely benefits of the proposed standard would be minimal. Other parties objected to the scope of the proposed standard, urging that milling and processing operations be excluded from coverage entirely because operational differences between grain elevators and grain-milling facilities made a single standard inappropriate. These commentators pointed to the fact that slower, smaller milling operations produce less dust, and that the feed additives and other processing features unique to milling have a fire-retardant effect that significantly reduces the risk of explosion. OSHA also received numerous comments urging it to exempt small elevator facilities, or “country elevators,” from the standard entirely or, at a minimum, to establish postponed compliance dates for such facilities. These participants generally relied upon data showing fewer fires and explosions in country elevators, and showing that small elevators would bear a compliance cost grossly disproportionate to the risk they created. Id. at 49,599. None of the participants in the rulemak-ing contended that housekeeping was unnecessary, although many challenged specific aspects of the housekeeping requirements. In particular, participants opposed the requirement (49 Fed.Reg. 1009 (1984)) that the housekeeping plan “minimize” dust accumulations, arguing that the language appeared to mandate expensive pneumatic dust-control systems. Some participants also faulted the action-level concept on the grounds that (1) the cost of compliance was not rationally related to any benefit, since even Vs-inch accumulations present a hazard; (2) the Vs-inch concept would create a false sense of security; and (3) it failed to direct attention where it was most needed — areas near and around known potential ignition sources. II. The Final Standard and Its Rationale. On December 31, 1987, OSHA promulgated a final standard for grain-handling facilities. 52 Fed.Reg. 49,592 (1987). The agency found that although industry had undertaken research and education initiatives since the 1977 disasters, explosions and fires continued to occur in grain facilities and, therefore, regulatory action was warranted. OSHA noted that in 1985 alone, 22 explosions had occurred, resulting in four deaths and 20 injuries. Id. at 49,-594-95. The standard addressed these hazards with a system of integrated safety practices, employee training, and engineering controls. OSHA estimated that full compliance with the standard would result in approximately a 75-percent reduction in grain fires and explosions, and a 50-pereent reduction in the severity of the remaining incidents. Id. at 49,622. Of the four elements necessary for fires and explosions in grain facilities — air, confinement, fuel (grain dust), and ignition— OSHA determined that control of only ignition sources and dust would be feasible. As OSHA’s expert, Dr. C.W. Kauffman, noted at the hearings: Air is ubiquitous and the scale of required inerting would be quite large.... [Confinement can be dealt with most easily during construction of a facility. We know what to do, we know why the explosions occur, and it’s simply a question of taking care of these ignition sources and the dust available. Id. at 49,595. Accordingly, the standard mandates a series of controls to reduce sources that might ignite a grain fire or explosion: Tramp metal and other foreign objects capable of producing friction sparks or machine defects must be eliminated from the grain stream at the receiving pit, and again before the grain enters the processing equipment. 29 C.F.R. § 1910.272(j), (m). Regular preventive maintenance is required for all grain-facility machinery, and the bucket elevator head and boot must be accessible for maintenance, inspection, repair, and cleanup. Grain dryers in elevators must be capable of automatic shutoff, and any new dryers must be located outside the grain elevator or be protected by an explosion suppression system or surrounded by fire-resistant walls. Id. § 1910.272(i), (p)(3). The bucket conveyor must be equipped with monitors that will shut down the bucket elevator in the event of malfunction. Id. § 1910.272(p)(5). In addition, no hot work — such as welding or torch-cutting — is permitted within any grain-handling facility except under the direct supervision of an employer representative or with a specially-issued permit. Id. § 1910.272(f). Jogging to free a choked leg is prohibited. Id. 1910.272(p)(l). Bearings inside elevator legs must be equipped with vibration or heat sensors. Id. § 1910.272(p)(4). Belts and lagging must have surface electrical resistance not to exceed 300 megohms. Id. § 1910.272(p)(2). Under the housekeeping provision of the final rule, 29 C.F.R. § 1910.272(i), fugitive grain dust accumulations must be carefully controlled. In all grain facilities — both mills and elevators — the employer is required to “develop and implement a written housekeeping program that establishes the frequency and method(s) determined best to reduce accumulations of fugitive grain dust on ledges, floors, equipment, and other exposed surfaces.” Id. § 1910.272(i)(l) (emphasis added). In addition, grain elevators must initiate appropriate cleaning measures whenever dust accumulations reach a depth of Vs inch in any of three priority housekeeping areas: (1) floor areas within 35 feet of inside bucket elevators; (2) floors of enclosed areas containing grinding equipment; and (3) floors of enclosed areas containing grain dryers located inside the facility. Id. § 1910.272(i)(2). As an alternative to complying with this Vs-inch action level, the employer may develop and implement a housekeeping program that demonstrably provides the “equivalent protection.” Id. § 1910.272(i)(2)(ii). The housekeeping component of the final standard contains several changes from the proposed standard; two are significant for this proceeding. First, the Vs-inch action level applies only in priority areas, not facility-wide as first proposed. OSHA agreed with those participants who argued that an objective action level was essential for effective and enforceable housekeeping, because it provides an objective measure of compliance. 52 Fed.Reg. 49,608-11 (1987). Without an objective measure, the new housekeeping provision might present enforcement problems similar to those that had been encountered under the old housekeeping standard (29 C.F.R. § 1910.22(a)(1)) or the general duty clause (29 U.S.C. § 654(a)(1)). 52 Fed.Reg. 49,611 (1987). However, OSHA was persuaded that the proposed 200-square-foot average was also susceptible to problems of measurement and enforcement. Id. at 49,610-11. Moreover, the agency found that the greatest fire and explosion potential exists in and near the leg. Id. at 49,609-10. The agency therefore determined that the action-level requirement should focus on “priority areas,” designated such because they contain a high concentration of ignition sources that cannot be effectively eliminated by the maintenance, engineering, and other controls required by the standard. Id. at 49,610. OSHA believed that by requiring cleanup whenever any accumulation in the priority area reaches Vs inch, the final action level would effectively reduce the total amount of grain dust in priority areas below what would have been allowed under the proposed rule. A second basis of departure from the proposed standard was to apply the Vs-inch action level to grain elevators only, thereby excluding grain mills. OSHA’s reasons, to be considered below in greater detail, were as follows: The evidence convinced OSHA that the vast majority of explosions in mills originate in the grinding equipment. Thus, OSHA reasoned, prevention of these primary explosions can be prevented by removal of foreign materials in the grain stream and by vigorous preventive maintenance. Id. at 49,613 (1987). OSHA also noted that mills are much cleaner than elevators because mills are subject to Food and Drug Administration (FDA) regulations requiring clean and sanitary operations where food and feed products are processed. Id. at 49,597, 49,613. See 21 C.F.R. § 110.37. Finally, OSHA found that the wide range of additives which feed manufacturers use, particularly limestone and liquids, greatly reduce the ability of a grain mixture to generate combustible dust. 52 Fed.Reg. 49,613 (1987). In another significant rulemaking determination, OSHA refused to exempt small elevators from the standard, concluding that the number of incidents, deaths, and injuries in small facilities, though lower than in large facilities, remained significant. OSHA found that the lower number of accidents did not reflect a reduced risk, since small elevators tend to operate, for the most part, only during harvest seasons. Id. at 49,598-99. OSHA was persuaded by participants, however, that very small elevators (less than one-million-bushel storage capacity) would have difficulty absorbing the costs of engineering controls and that their small size made alternatives to those controls feasible. Id. at 49,600. Thus, small elevator operators were permitted to substitute daily visual inspection both for installation of the motion detection devices and for belt alignment monitoring devices otherwise required for bucket elevator legs. See id. at 49,599-601, 49,619; 29 C.F.R. § 1910.272(p)(7). OSHA estimated the standard’s annual compliance cost at $5.7 million for mills, or less than one percent of after-tax profits. For elevators, the annual compliance cost was estimated at $35.7 to $63.1 million. Broken down for each segment of the industry, compliance estimates for elevators represented 1.56 percent (for inland terminals) to 7.10 percent (for export terminals) of net income under a once-per-shift calculation, and 2.73 percent (for inland terminals) to 21.10 percent (for export terminals) of net income under a three-passes-per-shift alternative. OSHA concluded that compliance would occasion no significant changes in the ratios of total liabilities to net worth, current assets to current liabilities, or net income to equity for elevators. III. Challenges to the Validity of the Standard. A. Scope of Review. The overall policy of the OSH Act is to “assure so far as possible ... safe'and healthful working conditions” for every American worker. 29 U.S.C. § 651(b). To that end, the Act authorizes OSHA to set “mandatory occupational safety and health standards,” which are defined by section 3(8) as “practices, means, methods, operations, or processes, reasonably necessary or appropriate to provide safe or healthful employment and places of employment.” 29 U.S.C. § 652(8). At least two provisions in the OSH Act limit OSHA’s power to set protective standards. The first is section 3(8)’s requirement that the standard be “reasonably necessary or appropriate” to protect employee health or safety. Id. To comply with this requirement, OSHA must make a threshold showing that a significant risk of either health impairment or job safety exists and that the risk can be eliminated or reduced by the change which the standard will impose upon industry practices. Industrial Union Dep’t, AFL-CIO v. American Pe troleum Inst., 448 U.S. 607, 641-42, 100 S.Ct. 2844, 2863-64, 65 L.Ed.2d 1010 (1980) (plurality opinion); Forging Indus. Ass’n v. Secretary of Labor, 773 F.2d 1436, 1445-46 (4th Cir.1985) (en banc). The second limitation is the Act’s “feasibility” requirement, contained in section 6(b)(5): The Secretary, in promulgating standards dealing with toxic materials or harmful physical agents under this subsection, shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.... In addition to the attainment of the highest degree of health and safety protection for the employee, other considerations shall be ... the feasibility of the standards. 29 U.S.C. § 655(b)(5) (emphasis added). Although section 3(8) does not contain an explicit feasibility requirement, the Supreme Court has stated that “any standard that [is] not economically or technologically feasible a fortiori would not be ‘reasonably necessary or appropriate’ under section 3(8).” American Textile Mfrs. Inst. v. Donovan, 452 U.S. 490, 513 n. 31, 101 S.Ct. 2478, 2492 n. 31, 69 L.Ed.2d 185 (1981). The feasibility requirement contained in section 6(b)(5)’s first sentence is not merely repetitive of the implied feasibility requirement of section 3(8). Section 6(b)(5) is an affirmative mandate as well as a limitation. Thereunder, the Secretary shall issue standards dealing with toxic materials or harmful physical agents to the extent feasible, that is, to the extent to which industry is “capable” of abating the hazard. American Textile, 452 U.S. at 508, 101 S.Ct. at 2490. Thus, “[a]ny standard based on a balancing of costs and benefits by the Secretary that strikes a different balance than that struck by Congress would be inconsistent with the command set forth by § 6(b)(5).” Id. at 509. However, OSHA standards promulgated under provisions other than section 6(b)(5) are subject to cost-benefit justification under section 3(8)’s “reasonably necessary” language. See Asbestos Information Ass’n/N. Am. v. OSHA, 727 F.2d 415, 423 (5th Cir.1984) (an emergency temporary standard “must, on balance, produce a benefit the costs of which are not unreasonable”). Section 6(f) of the OSH Act, 29 U.S.C. § 655(f), establishes substantial evidence as the standard for judicial review of OSHA’s standards. Under the substantial evidence test, “we must take a ‘harder look’ at OSHA’s action than we would if we were reviewing the action under the more deferential arbitrary and capricious standard applicable to agencies governed by the Administrative Procedure Act[, 5 U.S.C. 706].” Asbestos Information, 727 F.2d at 421 (footnote omitted). The substantial evidence test applies both to OSHA’s factual findings and to its essentially legislative policy decisions. Texas Indep. Ginners Ass’n v. Marshall, 630 F.2d 398, 404 (5th Cir.1980). For factual determinations, the Supreme Court has defined substantial evidence as “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 477, 71 S.Ct. 456, 459, 95 L.Ed. 456 (1951) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 216, 83 L.Ed. 126 (1938)). Our review of factual findings, therefore, is not de novo. Even if the evidence supports inconsistent inferences, we will not displace the agency’s choice so long as a reasonable person could reach that conclusion. Public Citizen Health Research Group v. Tyson, 796 F.2d 1479, 1485 (D.C.Cir.1986). For policy determinations, OSHA’s decision must be scrutinized even though, unlike factual findings, it is not susceptible to verification or refutation by the record. Texas Indep. Ginners, 630 F.2d at 404. This scrutiny involves a two-pronged analysis: The Secretary’s action must be found (1) consistent with the statutory language and purpose of the OSH Act, and (2) reasonable under the state of the record before her. Id. Where the agency’s policy determinations are based upon complex scientific and factual data, or involve speculative projections, our review is particularly deferential. Forging Indus. Ass’n, 773 F.2d at 1443. Armed with the foregoing principles, we review the union’s petitions first, then the industry’s. Their only commonality is challenge to the efficacy of the currently-mandated action level. The unions contend that under section 6(b)(5) of the OSH Act, the Agency is required to devise a grain-handling-facilities standard that affords grain workers the maximum protection feasible. Instead, they argue, OSHA gave industry’s compliance costs much more weight than does the balance struck by Congress. Specifically, the unions believe the final standard affords the employers discretion to provide only minimal protection. It requires all grain-handling facilities to develop their own written housekeeping plan, specifying the methods and frequency for “reducing” grain dust accumulation, but not “minimizing” it. No specific dust-control measures are required to be implemented. Nor is there any objective limit placed upon the amount of dust that may accumulate without violating the standard. The Vs-inch action level is not a specific limit on accumulations, but rather is merely a directive that an employer initiate clean-up operations whenever dust reaches the specified level. B. Union Challenges. The unions do not quarrel with the action-level concept, but would prefer it be implemented on a facility-wide basis. This would more effectively reduce the amount of dust available to fuel secondary explosions, in addition to controlling ignition in priority areas. The unions also argue that OSHA should impose a facility-wide action level in grain mills as well as grain elevators, or demonstrate why such a measure would not be feasible. 1. The Meaning of “Harmful Physical Agents” Under Section 6(b)(5) of the OSH Act. As a preliminary matter, we must decide whether an OSHA safety standard, regulating the incidence of physical injury from an immediate and obvious danger, must comply with the feasibility mandate contained in section 6(b)(5) of the OSH Act. All parties look to the Supreme Court’s decision in American Textile Mfrs. Inst., Inc. v. Donovan as controlling authority. In that case, a coalition of industry groups challenged OSHA’s cotton-dust standard, which was promulgated to protect workers from byssinosis, a progressive respiratory disease caused primarily by the inhalation of cotton dust. Industry challenged the standard under both sections 3(8) and 6(b)(5) of the OSH Act, contending that the standard was neither reasonably necessary to protect workers, nor technologically or economically feasible to implement. The District of Columbia Circuit sustained the standard in most major respects in AFL-CIO v. Marshall, 617 F.2d 636 (D.C.Cir.1979), aff'd in relevant part sub nom. American Textile Mfrs. Inst. v. Donovan, 452 U.S. 490, 101 S.Ct. 2478, 69 L.Ed.2d 185 (1981). In so holding, that circuit rejected the approach taken by this circuit in American Petroleum Inst. v. OSHA, 581 F.2d 493 (5th Cir.1978), aff'd on other grounds sub nom. Industrial Union Dep’t v. Am. Petroleum Inst., 448 U.S. 607, 100 S.Ct. 2844, 65 L.Ed.2d 1010 (1980), which had required OSHA to conduct a cost-benefit analysis in promulgating toxic substance standards under section 6(b)(5). On writ of certiorari, the industrial petitioners placed the cost-benefit issue squarely before the Court, asserting that section 6(b)(5)’s “to the extent feasible” language necessarily implies some form of cost-benefit justification. The Court disagreed, holding that OSHA need not, indeed must not, consider cost-benefit criteria in setting toxic-substance standards. 452 U.S. at 506-22, 101 S.Ct. at 2489-97. See also Building & Constr. Trades Dep’t, AFL-CIO v. Brock, 838 F.2d 1258, 1264 (D.C.Cir.1988) (“[T]he [American Textile ] Court held that the ‘feasibility’ standard of § 6(b)(5) does not require the Secretary to balance cost and benefit in defining a standard, and clearly manifested the Court’s belief that the [OSH] Act did not permit the Secretary to do so.”) (emphasis in original). To reach this result, the Court employed a two-step analysis. First, it defined the statutory term “feasible” according to its common dictionary meaning — that which is “capable of being done, executed, or effected.” 452 U.S. at 508-09, 101 S.Ct. at 2490-91 (quoting Webster’s Third New Int’l Dictionary of the English Language at 831 (1976)). Thus, the Court reasoned that § 6(b)(5) directs the Secretary to issue the standard that ‘most adequately assures ... that no employee will suffer material impairment of health,’ limited only by the extent to which this is ‘capable of being done.’ In effect then, as the Court of Appeals held, Congress itself defined the basic relationship between costs and benefits, by placing the ‘benefit’ of worker health above all other considerations save those making attainment of this ‘benefit’ unachievable. Any standard based on a balancing of costs and benefits by the Secretary that strikes a different balance than that struck by Congress would be inconsistent with the command set forth in § 6(b)(5). Thus, cost-benefit analysis by OSHA is not required by the statute because feasibility analysis is. Id. at 509, 101 S.Ct. at 2490 (footnote omitted, emphasis added). The Court next turned to the question of whether section 3(8)’s requirement that a standard be “reasonably necessary or appropriate” should be construed to impose cost-benefit constraints upon OSHA’s rulemaking authority. Assuming arguendo that section 3(8) implies a cost-benefit justification, the Court nonetheless refused to read that requirement as overriding section 6(b)(5)’s feasibility mandate for standards that regulate toxic substances and harmful physical agents. The Court reasoned that to hold otherwise would effectively eviscerate section 6(b)(5)’s “to the extent feasible” requirement: The level of risk deemed acceptable under cost-benefit criteria would never be lower — but rather, would typically be higher — than the level dictated by feasibility analysis. 452 U.S. at 513, 101 S.Ct. at 2492. Thus, where OSHA sets standards under section 6(b)(5), that section’s feasibility mandate preempts any criteria, less protective of workers, that might be read into section 3(8). One question expressly left open by the Court in American Textile is whether section 3(8) itself implies a cost-benefit justification for standards promulgated under enabling provisions other than section 6(b)(5). 452 U.S. at 513 n. 32, 101 S.Ct. at 2493 n. 32. Accordingly, OSHA presently contends that the rule before us is such a standard, since grain dust is neither a “toxic substance” nor a “harmful physical agent” under section 6(b)(5). In making this argument, OSHA urges that the final clause of section 6(b)(5)’s first sentence makes abundantly clear that the section’s coverage is limited to substances that take their toll over time or whose “deleterious effect is not readily apparent,” such as a carcinogen or a harmful agent such as noise. The Secretary points to critical language in that clause requiring her to issue the standard designed to assure that “no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.” 29 U.S.C. § 655(b)(5) (emphasis added). The unions assert that OSHA’s argument in plainly at odds with the plain text of section 9(b)(5): They assert that OSHA would have us rewrite the phrase “even if such employee has regular exposure to the hazard ... for the period if his working life” by eliminating the words “even if,” a term of inclusion, and substituting in their place the phrase “only if,” a term of exclusion. The Secretary, on the other hand, explains that the “even if” clause serves the purpose, not simply of adding a category of hazards with which the Secretary should be concerned, but of directing that when standards “dealing with toxic materials or harmful physical agents” are promulgated, they should be set at a level sufficiently low to protect all workers from the latent effects of exposure to such materials or agents, a factor not generally relevant to safety standards. There are Supreme Court pronouncements that can be read to support such a distinction. In Industrial Union Dep't AFL-CIO v. American Petroleum Inst., a plurality of the Court asserted, in a footnote, that [t]he reason that Congress drafted a special section for [toxic materials and harmful physical agents] was not ... because it was thought that there was a need for special protection in these areas. Rather, it was because Congress recognized that there were special problems in regulating health risks as opposed to safety risks. In the latter case, the risks are generally immediate and obvious, while in the former, the risks may not be evident until a worker has been exposed for long periods of time to particular substances. It was to ensure that the secretary took account of these long-term risks that Congress enacted § 6(b)(5). 448 U.S. at 649 n. 54, 100 S.Ct. at 2867 n. 54 (emphasis added). Citing this footnote, the Court later stated in American Textile that “Congress could reasonably have concluded that health standards should be subject to different criteria than safety standards because of the special problems presented in regulating them.” 452 U.S. at 512, 101 S.Ct. at 2492 (emphasis in original). Whether Congress did in fact reach such a conclusion is not as clear as it might be in the legislative history, however. The legislative history upon which the Industrial Union plurality based its footnote was a single statement by Senator Dominick. This statement, moreover, like the first sentence of section 6(b)(5) itself, does not evince Congress’ intention to exclude risks of immediate danger, but to include hazards occasioned by long-term exposure. The senator explained that Congress’ intention in adopting section 6(b)(5) was to require the Secretary to use his best efforts to promulgate the best available standards, and in so doing ... he should take into account that anyone working in toxic agents and physical agents which might be harmful may be subjected to such conditions for the rest of his working life, so that we can get at something which might not be toxic now, if he works in it a short time, but if he works in it the rest of his life might be very dangerous; and we want to make sure that such things are taken into consideration in establishing standards. 116 Cong.Reg. at 37,622-23 (quoted in Industrial Union, 448 U.S. at 648-49, 100 S.Ct. at 2867-68) (emphasis added). Returning once again to the OSH Act itself, we note that Congress made no blatant effort to define or distinguish between health and safety standards. On the contrary, the third sentence of section 6(b)(5), which enumerates policy considerations for all section 6(b)(5) standards, arguably rejects such a distinction. It provides that in addition to the attainment of the highest degree of health and safety protection for the employee other considerations shall be the latest available scientific data in the field, the feasibility of the standards, and experience gained under this and other health and safety laws. 29 U.S.C. § 655(b)(5) (emphasis added). The sharp distinction apparently drawn in American Textile between health and safety standards is undercut, moreover, by statements elsewhere in the opinion that seemingly ignore it. For example, in discussing the feasibility requirement, the Court states that “[njowhere is there any indication that Congress contemplated a different balancing by OSHA of the benefits of worker health and safety against the costs of achieving them.” 452 U.S. at 521, 101 S.Ct. at 2496 (emphasis added). To support this statement, the Court cited congressional concern to protect workers from all types of occupational health and safety standards. Id. at 522, 101 S.Ct. at 2497. The Court thus concluded that Congress intended employers to bear substantial costs “when necessary to create a safe and healthful working environment. Congress viewed the costs of health and safety as a cost of doing business.” Id. at 521, 101 S.Ct. at 2496 (emphasis added; footnote omitted). These statements, and the use of congressional authority, offer some support for the unions’ view that Congress was concerned with protecting workers equally from both safety and health hazards, not merely the latter. We are faced, therefore, with a statute and legislative history that are not as clear as they could be and with arguable inconsistent language from Supreme Court opinions. Our role in reviewing a regulation under such circumstances is set forth in Chevron, U.S.A., Inc. v. Natural Re sources Defense Council, Inc., 467 U.S. 837, 844-45, 104 S.Ct. 2778, 2782-83, 81 L.Ed.2d 694 (1984): We have long recognized that considerable weight should be accorded to an executive department’s construction of a statutory scheme it is entrusted to administer, and the principle of deference to administrative interpretations ‘has been consistently followed by this Court whenever decision as to the meaning or reach of a statute has involved reconciling conflicting policies, and a full understanding of the force of the statutory policy in the given situation has depended upon more than ordinary knowledge respecting the matters subjected to agency regulations. ‘If this choice represents a reasonable accommodation of conflicting policies that were committed to the agency’s care by the statute, we should not disturb it unless it appears from the statute or its legislative history that the accommodation is not one that Congress would have sanctioned.’ United States v. Shimer, 367 U.S. 374, 382, 383, 81 S.Ct. 1554, 1560, 1561, 6 L.Ed.2d 908 (1961). (Footnote and citations omitted). Thus, only where congressional intent is pellucid are we entitled to reject reasonable administrative construction of a statute. INS v. Cardoza-Fonseca, 480 U.S. 421, 447-48, 107 S.Ct. 1207, 1221-22, 94 L.Ed.2d 434 (1987). Where, on the other hand, an agency’s interpretation is a reasonable reading of intent, we defer to it. See Brock v. Schwarz-Jordan, Inc., 777 F.2d 195, 198 (5th Cir.1985) (per curiam). Were we to divine Congress’s intent de novo in this matter, we might with some ease conclude that the better view is that Congress, if faced squarely with the issue, would have required OSHA to develop standards to assure worker safety equally as to both latent health hazards and hazards, such as explosives, that are every bit as lethal but whose impact is immediate. However, given the ambiguity visited here by Congress, and in the absence of a specific exposition by the Supreme Court, we defer to OSHA. Accordingly, we hold that the immediate and obvious danger posed by grain dust in grain-handling facilities does not constitute a “harmful physical agent” within the contemplation of section 6(b)(5). OSHA thus was not subject to the feasibility mandate contained in that section in promulgating a standard to reduce or eliminate that danger. Hence, the appropriate standard for OSHA is that its standard be “reasonably necessary or appropriate” to protect employee safety under section 3(8). This determination encompasses a specie of cost-benefit justification, as we explained in Asbestos Information Ass’n/N. Am. v. OSHA, 727 F.2d [415] at 423 [(5th Cir.1984)]. The test under section 3(8) is an intermediate one between the feasibility mandate of section 6(b)(5) and a strict cost-benefit analysis that requires a more formal, specific weighing of quantified benefits against costs. Thus, in Asbestos we noted that an emergency temporary standard must, on balance, produce a benefit the costs of which are not unreasonable.... [W]e do not imply that [the OSH Act] requires the agency to [conduct a formal cost-benefit analysis] before it promulgates [such a standard].... We cannot say that the cost of compliance is unreasonable if the [standard] in fact alleviates a grave danger. Id. at 423-24 & n. 18. Similarly, in Texas Indep. Ginners [v. Marshall ] we concluded that [t]he reasonably necessary requirement in [section 3(8)] only demands that the expected costs of OSHA regulations be reasonably related to the expected benefits, leaving considerable discretion for the agency as long as it is exercised on substantial evidence and with an adequate statement of reasons. 630 F.2d [398] at 411 n. 44 [(5th Cir. 1980) ]. 2. Scope of the Action Level: Priority Area or Facility-Wide? OSHA must present a statement of reasons for its action, see 29 U.S.C. § 655(e), and, as we have noted, “[i]t is axiomatic that the reasons the Agency gives at the time it acts form the basis for the Agency’s action.” Asbestos Information Ass’n, 727 F.2d at 425. However, because of our deferential review of OSHA’s legislative-type policy decisions, we do not require the agency to respond in detail to every imaginable proposal for tighter standards. Instead, a party challenging an OSHA standard bears the burden of demonstrating that the agency’s proposed standard is not “reasonably necessary or appropriate” to address the problem at hand. OSHA here recognizes that dust accumulations of Vs inch are “by no means a ‘safe’ level for grain dust.” 52 Fed.Reg. 49,611 (1987). The record is also replete with evidence that it is the secondary explosions that cause most of the deaths, injuries, and devastation, and that by controlling dust accumulations, the risk and magnitude of secondary explosions will be reduced. Id. at 49,607. OSHA further acknowledges that potential ignition sources in elevators “cannot be readily identified and controlled with any certainty or comprehensiveness.” Id. at 49,611. Nonetheless, the agency confines the Vs-inch action level to areas where ignition of grain dust, and hence primary explosions, are “most likely” to occur. Presumably, in areas where ignition is less likely, primary explosions will continue to occur. And when they do, the level of dust available throughout the facility to fuel the secondary explosion will depend, in large part, upon the initiative taken by the employer to minimize dust accumulations pursuant to its own housekeeping plan. As discussed, OSHA ultimately rejected its own initial proposal that would have imposed a Vs-inch action level over any 200-square-foot area in the entire facility. This proposal was rejected largely for reasons of administrative efficiency, although the agency made passing mention that the approach “could result in problems ... with economic feasibility_” Id. at 49,-611. OSHA contends that the final standard embodies the optimal enforcement strategy because the boundaries of the priority areas are certain: If Vs inch has accumulated anywhere in the priority area, and the employer has not begun to remove the accumulation or reduce its combustibility, a citation may be issued. By contrast, the effectiveness of the 200-square-foot average was questionable, because of the self-evident uncertainty as to the boundaries of the area. OSHA also concluded that the facility-wide scope of the proposal’s action level, though superficially appealing, lacked focus on the most dangerous areas — the leg and dryer areas — thus undercutting the efficacy of the action level. The final action level requires clean-up whenever Vs inch accumulates in any part of the priority area, regardless of total amount. The proposal, on the other hand, would not have required cleanup until sufficient dust had accumulated to cover a 200-square-foot area to a Vs-inch average depth. Thus, in the most hazardous priority areas, the final standard tolerates far less dust than would have been tolerated under the proposal and, during active periods, requires much more diligent housekeeping. See id. at 49,610. Although OSHA’s reasons convince us of the administrative unmanageability of a facility-wide action level based upon dust accumulation averaged over a 200-square-foot area, these enforcement problems could be circumvented by simply extending a strict action level to the entire facility. By extending, in effect, the priority areas to encompass the entire floor space of the facility, the potential boundary disputes evaporate. More importantly, a specific, objective criterion that mandates removal of dust throughout the entire facility presumably would both reduce the risk of ignition, as envisioned under the current standard, and reduce the amount of dust available to fuel a secondary explosion, should ignition occur. As for technical considerations, evidently a facility-wide standard would require no more than a push broom. And as for cost, OSHA’s preliminary estimate was $24 million for a facility-wide action level — clearly within the range of estimated net compliance costs under the current proposed standard ($5.9 million to $33.4 million). Id. at 49,622. These factors lead us to conclude that extending the Vs-inch action level to the entire facility (without a 200-square-foot limitation) should have been considered by the Secretary. See Motor Vehicle Mfrs. Ass’n of the United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866-67, 77 L.Ed.2d 443 (1983); Public Citizen Health Research Group v. Tyson, 796 F.2d 1479, 1507 (D.C. Cir.1986); 29 U.S.C. § 655(e). Accordingly, we remand for further consideration of this question. On remand, OSHA must consider, as an option, the adoption of a facility-wide standard. Of course, it must also explain its decision to utilize, or to reject, such a standard. 3. Scope of the Action Level: Should Grain Mills Also Be Made To Comply? As mentioned above, OSHA imposed an action level for grain elevators but not for grain mills. For mills, OSHA believed it “possible to prevent primary explosions without imposing specific dust-accumulation limits in priority housekeeping areas, because the potential ignition sources can be identified and controlled.” Id. at 49,613 (emphasis added). Unlike grainelevator explosions, whose ignition sources OSHA concedes to be elusive, the vast majority of explosions in mills can be traced almost exclusively to the grinding equipment used in the facilities, and to the presence of tramp metal and other foreign materials, such as tools, which somehow enter the grain stream. The standard, therefore, controls these potential ignition sources by a combination of foreign-object screens, general housekeeping, employee training, and preventive maintenance. OSHA also points out that various differences between mill and grain elevator operations make mills less susceptible to fires and explosions. Mills typically operate at slower speeds and have smaller capacities than grain elevators. Because mills also generally run year-round with little seasonal change, maintenance of equipment is more systematic. Moreover, mills typically use a myriad of ingredients which, when mixed with grain, produce a significantly less flammable substance than raw grain dust. Finally, grain mills are already subject to stringent housekeeping and infestation standards enforced by other agencies. Since these regulations are concerned both with keeping the feed from becoming adulterated by the work environment and with preventing cross contamination between batches of product, OSHA concludes that the cleaning procedures thus entailed will reduce the likelihood that grain dust will be allowed to accumulate anywhere in the facilities. Id. The unions rely upon United Steelworkers of Am. v. Auchter, 763 F.2d 728 (3d Cir.1985), on motion to enforce judgment, 819 F.2d 1263 (3d Cir.1987), in arguing that OSHA improperly excluded grain mills from compliance with an action level comparable to that imposed upon grain elevators. In that case, OSHA found that all workers face a significant risk from the lack of information about chemical hazards, but that the risk is greater among workers in the manufacturing sector. For this reason, OSHA limited coverage of its hazard-communication standard to the manufacturing sector, excluding employees in such sectors as service, construction, and agriculture. Significantly, the record showed that some workers in those non-covered sectors had higher reported rates of chemical-source illness and injury than did workers in many covered industries. The court therefore remanded the standard for further consideration by the agency, holding that [o]nce a standard has been promulgated ... the Secretary may exclude a particular industry only if [s]he informs the reviewing court, not merely that the sector selected for coverage presents greater hazards, but also why it is not feasible for the same standard to be applied in other sectors where workers are exposed to similar hazards. 763 F.2d at 738. The unions accordingly contend that here, as in United Steelworkers, OSHA found that all grain industry workers face a similar hazard — i.e., physical injury from fire and explosions — but that those risks are greater among workers in a particular sector, grain elevators. Thus, the unions reason that OSHA should extend mill workers the same coverage, or explain why doing so would not be feasible. Because we find United Steelworkers distinguishable from the case before us, we reject the unions’ argument. Whereas in that case, OSHA totally excluded coverage to workers it had determined were at significant risk, here, by contrast, OSHA has afforded mill workers coverage that purportedly will eliminate the significant risk of fires and explosions in mills. The burden, therefore, remains on the unions to demonstrate that their proposal will both provide more than a de minimis benefit for mill worker safety and be feasible to implement. Since the unions have shown neither, we do not find that the Secretary has abused her discretion in refusing to mandate an action level for grain mills. C. Industry Challenges. The industry contends that OSHA’s grain-handling-facilities standard is based upon erroneous cost-benefit calculations. We need not reach this argument, already having decided that section 6(b)(5) of the OSH Act preempts it. Once again, as the American Textile court observed, “cost-benefit analysis is not required by the statute because feasibility analysis by OSHA is.” 452 U.S. at 509, 101 S.Ct. at 2490 (footnote omitted). Otherwise, the industry raises two specific challenges: First, it argues that OSHA has failed to make the threshold showing, required under section 3(8) of the OSH Act, that the action-level component of its final standard will eliminate or reduce a significant risk. Second, it argues that the action level, in any event, is not economically feasible. We now address each of the industry’s arguments in turn. 1. Reduction of Significant Risk. Like the unions, industry builds upon OSHA’s concession that Vs-inch is not a safe level for grain dust. If this is so, industry reasons, then what does a Vs-inch action level hope to accomplish? Industry points to the findings of the Factory Mutual Study, which demonstrated that dust accumulations (test medium was corn starch) of as little as Vioo inch could fuel the spread of a flame front. This level, however, is unachievable, and industry thus concludes that OSHA should not set an action level at all. We are not persuaded by this argument. We find substantial evidence in the record that OSHA’s choice of a Vs-inch action level for priority areas, although unable to eliminate the significant risk of grain elevator explosions, will substantially reduce that risk. See 52 Fed.Reg. 49,612 (1987). That is all section 3(8) requires. Studies performed by USDA, NAS, NIOSH, and Kauffman and Hubbard uniformly agree that dust is fuel for fire and explosion and should be eliminated to the extent possible. The high concentration of uncontrollable ignition sources in and around grain elevator bucket legs provides abundant support for OSHA’s conclusion that an especially intensive housekeeping effort is warranted for those priority areas. Contrary to the industry’s assertions, the record shows that the amount of layered dust is directly related to the risk in several respects. First, the greater the depth of layered dust in any enclosed space, the greater is the likelihood that it can produce ambient concentrations at explosive levels. Second, the greater the area within the facility covered with layered dust, the more likely it is that a confined space will be involved. Third, the more areas compromised by layered dust, the greater is the likelihood that a fire, primary explosion, or secondary explosion will harm an employee. And finally, the greater the ambient concentration of dust, the more severe is the explosion, because the total energy produced by combustion reflects the amount of available fuel. To be sure, though one study cited by the industry does show that there is an upper limit on the amount of pressure (energy) which layered dust can produce when caused to mix with air and an ignition source, that finding does not nullify the very real relationship between the amount of layered dust and explosive energy below that ceiling. OSHA need not await the “Godot of scientific certainty” while allowing workers to suffer. United Steelworkers of Amer. v. Marshall, 647 F.2d at 1266. Based upon all of these considerations, OSHA concluded that Vs-inch was an appropriate limit because it is visible, measurable, and removable, and that the Vs-inch action level provides the objective measure of compliance necessary for effective guidance to the industry and effective enforcement by the agency. Finding this determination to be supported by substantial evidence, we will not displace the agency’s choice. 2. Economic Feasibility. A standard is economically feasible if industry’s “long-term profitability and competitiveness” is not threatened. American Textile, 452 U.S. at 531 n. 55, 101 S.Ct. at 2501 n. 55. In reviewing OSHA’s determination that compliance with the standard will not cause “serious economic dislocation” in entire industries, see AFL-CIO v. Brennan, 530 F.2d 109, 123 (3d Cir.1975), we “do not review [the agency’s] cost figures de novo, but accord [the agency] discretion to arrive at a cost figure within a broad zone of reasonable estimate.” Forging Indus. Ass’n v. Secretary of Labor, 773 F.2d 1436, 1453 (4th Cir.1985) (en banc) (quoting Weyerhaeuser v. Costle, 590 F.2d 1011, 1049 (D.C.Cir.1978)). The relevant inquiry, then, is “whether OSHA has constructed ‘a reasonable estimate of compliance costs and demonstrate[d] a reasonable likelihood that these costs will not threaten the existence or competitive structure of an industry, even if it does portend disaster for some marginal firms.’ ” Id. (quoting United Steelworkers, 647 F.2d at 1272). For its grain-handling-facilities standard, OSHA predicted that compliance costs could move as many as 183 grain elevators into negative net income. However, since this figure represents less than one percent of all covered elevators, OSHA found the standard to be feasible. The industry does not challenge OSHA’s conclusion as a matter of policy, but challenges, rather, the facts upon which it is based. Since the labor cost in complying with the action-level provision is the single most expensive item of compliance for grain elevators, the rate at which priority areas can be swept and vacuumed is critical to OSHA’s prediction. The industry contends that OSHA significantly (by a factor of 5.4 to 1) understated the cost of compliance for grain elevators by applying an incorrect vacuuming and sweeping rate of 8,100 square feet per hour, and asserts that the proper cleaning rate is only 1,000 to 1,500 square feet per hour. OSHA based its estimate of housekeeping labor costs upon two studies of vacuuming rates: A time-and-motion study of vacuuming in a factory setting, and an American National Standards Institute (“ANSI”) study of carpet-sweeping. The time-and-motion study showed a vacuuming rate of 9,400 to 11,000 square feet per hour over an uncarpeted surface. The ANSI report, applying standard test methodology for relative work, showed a vacuuming rate of 6,480 square feet per hour over a carpeted surface with a 12-inch vacuum head. This translate