Industrial Union Dept., AFL-CIO v. American Petroleum Institute

DEPARTMENT, AFL-CIO UNION INDUSTRIAL v. INSTITUTE PETROLEUM AMERICAN et al. 2, July 1980* 1979 Decided Argued October 78-911. No. American v. Labor 78-1036, Marshall. No. *Together with court. to the same al., on certiorari also et Institute Petroleum J., announced Stevens, judgment of the Court- and delivered an opinion, in which Burger, J.,C. J., joined, I, Stewart, and in Parts II, III-A, III-B, III-C, and III-E of J., which joined. Burger, Powell, J.,C. filed a concurring opinion, post, p. 662. J., opinion filed an Powell, concurring part and concurring judgment, post, p. 664. Rehn- quist, J., opinion filed an concurring judgment, post, p. 671. Mar- J., filed a dissenting opinion, in shall, which Brennan, and White, JJ., joined, post, p. 688. Blackmun, George H. Cohen argued the petitioner cause for in No. 78-911. With him on the briefs were Robert Weinberg, M. Woll, J. Albert Gold, Laurence Elliot Bredhojf, George and William Alsup argued the cause for petitioner Kaufmann. in No. 78-1036. With him on the briefs were Solicitor Gen- McCree, eral Deputy Solicitor Easterbrook, General Ben- jamin Mintz, W. K. Dennis and Kade.

Edward W. Warren argued respondents cause for Amer ican Petroleum Institute et inal. both cases. him With on the Ritchie, were Stark brief Martha Beauchamp, King, Neil J. Pickering, Bonczek, John H. Robert R. John F. Dickey, Ackerly, Robert Harold B. Scoggins, L. and Jr. Charles F. argued Lettow respondents cause for Rubber Manufac turers Inc., et in both Association, al. cases. With him on the † brief was John Jr. Murphy, C. judgment announced the Court

Me. Justice Stevens opinion, and delivered an in which The Chief and Justice joined III- I, II, III-A, and in Parts Justice Stewaet Me. III-C, III-E which B, joined. and Justice Powell Me. Occupational Safety Act of 1970 (Act), and Health seq., pur- Stat. S. C. et was enacted for the U. pose working and for ensuring safe healthful conditions every working man and woman in the This litiga- Nation. tion promulgated by concerns a standard regulate occupational exposure benzene, Labor to a sub- high expo- been stance which has shown to cause cancer at principal sure levels. The question is whether such show- for the most ing places a sufficient basis a standard that stringent techno- limitation to benzene that logically economically possible. and Secretary delegates authority

The Act broad to the promulgate different kinds of The basic definition standards. FiUion for of amid curiae John A. urging by reversal were filed

† Briefs Agricultural Union, Aerospace and Automobile, the International United Ayres E. Richard America; by for the Natural Implement Workers Council, Defense Inc. Resources V. J. by Briefs of amici curiae urging affirmance were filed Alfred Yohay by Anthony Stephen C. J. Obadal Co.;

Prather for Anaconda Zener, Stephen A. Robert V. Foundation; by Capital Legal for Bokat, L. Kovacs for and William of Commerce the Chamber United States. Holt, Kilberg, Thaddeus William J. Briefs of amici curiae were filed B. Robinson for Inc.; by Lawrence Z. David Lorber ASARCO James R. Richards by Association; Manufacturers and the Chocolate Joseph et al. Cimino “occupational safety and health standard” is found in § 3 (8), which provides:

“The term ‘occupational safety and health standard’ means a standard which requires conditions, adop- or the tion or use of one or more practices, means, methods, operations, or processes, reasonably necessary or appro- priate provide safe or employment healthful places and of employment.” 1591, Stat. U. S. 652§C. Where toxic physical materials harmful agents con- are cerned, a standard must also comply with § 6 (b)(5), which provides:

“The Secretary, in promulgating standards with dealing toxic materials or physical harmful agents under this subsection, shall set the standard which adequately most to the .extent assures, feasible, basis the best available evidence, no employee will suffer material impairment of health or functional capacity even if such employee regular has exposure to the hazard dealt such standard for period of his working life. Devel- opment of standards under this subsection shall be based upon research, demonstrations, and experiments, such other information may be appropriate. In addition attainment of highest degree of health and safety protection for the employee, other considerations shall be the latest available scientific data in the field, feasibility standards, and experience gained under and other health safety and laws.” 84 Stat. 1594, 29 S. C. § U. (b)(5).1 1The second third and sentences of this section, which impose feasibility on the limits him allow to take into account the best avail developing standards, may able evidence in apply to all health safety if standards. This conclusion follows the term “subsection” used in the to the second sentence refers entire subsection (b) (which sets out procedures adoption types for the all of health safety standards), simply to the toxic rather than materials subsection, (b)(5). While §6 post, Marshall, respondents Me. Justice agree with this

613 regulated the material a carcinogen, Wherever toxic is Secretary position has taken the exposure the that no safe can (b) (5) requires level be determined and 6 § him limit exposure technologically to set an at the lowest feasible reg- not impair viability level that will of the industries In having this after determined there is a case, ulated. (a between benzene cancer causal connection and leukemia limit exposure cells), of the white blood set per benzene part of one on airborne concentrations benzene eye con- (1 regulated dermal and parts ppm), million of air complex imposed and containing with solutions tact benzene, employers monitoring testing requirements medical 29 of benzene. ppm or more workplaces contain 0.5 whose (1979). (e) 1910.1028 (c), §§ CFR §C. 655 29 S. pursuant review U. pre-enforcement On Fifth Circuit for Appeals (f), the United States Court Institute Petroleum American invalid. regulation held court concluded OSHA, (1978). 2d 493 581 F. v. 2 Safety Health Administration Occupational (QSHA) it had authority because standard-setting exceeded its had “reason- limit new benzene not shown that or healthful safe appropriate provide ably necessary or (b)(5) 6§ (8),3 3 and because required employment” et al. Institute Respondents Petroleum American position, see Brief for 1137, 1107, OSHA, J. Currie, Am. Bar Found. Research 39; 1976 see also 58; see also not, for Federal Parties 151, Brief does see n. Government Feasibility Regulating Technological Berger Riskin, Economic and & 7Act, Ecol- Safety Health Occupational Toxic Substances Under this issue us to decide need for ogy There is no L. Q. these cases. of Labor Department agency within the administrative OSHA is the the Act. enforcing standards under promulgating and responsible “Agency” “Secretary,” “OSHA” opinion, refer to the we this interchangeably. only standards obligation to enact imposes on OSHA Act “The healthful provide safe or necessary appropriate reasonably that are definition, one not fit in does workplaces. If a standard 2d, F. at 502. to enact.” is authorized that OSHA does give “not OSHA the unbridled discretion to adopt stand- ards designed to create absolutely risk-free workplaces re- gardless of costs.” Reading the two provisions together, Fifth Circuit held that the Secretary duty was under a determine whether the expected benefits from the stand- new ard bore a reasonable relationship to the costs im- that Id., at 503. The court noted that OSHA had made posed. an estimate of the costs of compliance, but that the record lacked substantial of any evidence discernible benefits.5 agree

We with the’ Fifth Circuit's holding (8)3§ that requires the Secretary find, to as a threshold matter, the 4 “Although 29 (b) S.U. C. (5) requires A. 655 goal § of attaining highest degree safety health and protection for the employee, it give does not OSHA the unbridled adopt discretion to designed standards to create absolutely workplaces risk-free regardless of Cost. To the con trary, requires section standards to feasible, and it contains a number pragmatic limitations in the specific form of kinds of infor mation OSHA must consider in enacting dealing standards with toxic materials. Those include 'the best evidence,’ available 'research, demon strations, experiments, and such other may information as be appropriate,’ ‘the latest available scientific field,’ data in the 'experience and gained under this and other safety health and Moreover, laws.’ dealing standards with toxic materials, just as with all occupational other safety and health standards, the conditions requirements other and imposed by the standard must be ‘reasonably necessary appropriate or provide to safe or healthful employment places and employment.’ (8).” U. S. C. A. 652 Ibid. 5“The lack of substantial evidence of discemable highlighted benefits is when one considers that OSHA point is unable to any empirical evidence documenting a leukemia risk at 10 ppm though even that has been the permissible exposure limit since 1971. OSHA’sassertion that benefits from reducing permissible exposure limit ppm from 10 1 ppm likely are to be appreciable, assumption only based on inferences drawn from involving studies higher exposure much levels rather than on studies in volving these or levels sound projections statistical high-level from the studies, satisfy does not reasonably necessary requirement limiting Aqua OSHA’s requires action. Slide OSHA to estimate the extent of expected benefits in order to determine whether those benefits bear a relationship reasonable to the standard’s demonstrably high Id., costs.” at 503-504. risk health significant a poses question substance toxic therefore lower standard new, a and workplace or safe provide appropriate necessary or “reasonably Unless^ employment.” places employment healthful necessary to address made, it finding is such until correctly Appeals Court whether question further costs/ between correlation a reasonable must there held Sec- argue, parties the federal whether, as benefits, promulgate (b)(5) by §6 required then retary is economically possible technologically as far goes risk. eliminate impres- first cases important unusually are these Because In this care. special record reviewed have sion, we analyze the (2) standard, the benzene (1) describe we opinion, limit, ppm1a imposing rationale Agency’s comment (4) issues, and legal controlling discuss (3) limitation. contact dermal briefly on

I It is commodity. important a familiar is Benzene under rapidly evaporates liquid colorless, aromatic a 11 billion Approximately conditions. atmospheric ordinary in States the United in produced were of benzene pounds by the produced total of that percent Ninety-four 1976. remainder industries, with petrochemical petroleum coking of byproduct industry as steel by produced of variety manufacturing in used is Benzene operations. much contain may (which fuels motor including products other pesticides, detergents, solvents, benzene), as 2% Reg. Fed. chemicals. organic exposed States the United of population entire The per parts fewa ranging benzene, quantities small Over 1029-1032. Tr. air. ambient in ppm, 0.5 billion ex- low-level additional subject are workers million one majority employment. their consequence aas posures benzene stations, service in gasoline work employees these production (petroleum refineries and coking operations), chemical processing, benzene transportation, rubber manu- facturing, and laboratory operations.6

Benzene is a toxic substance. Although it could con- ceivably cause harm to a person who swallowed or touched it, the principal risk of harm comes from inhalation of benzene vapors. When vapors these are inhaled, benzene diffuses through the lungs and quickly absorbed into the blood. figures OSHA’s indicate 795,000 service station employees have heightened some exposure to benzene as a result their employment. See 2 U. Dept, S. Labor, OSHA, Technology Assessment and Economic Impact Study of an Regulation OSHA for Benzene, p. 1977) D-7 (May (hereinafter Economic Impact Statement), 11 Record, 5B, Ex. p. D-7. These employees are specifically excluded from the regulation at issue in infra, this case. See at 628. OSHA states 629,000 that another employees, who are covered the regulation, work in the other industries described. Reg. 43 Fed. 5935 (1978). It is not clear from the record its explanation or permanent standard how OSHA arrived at the 629,000 estimate of exposed employees. OSHA’s consultant, Arthur Little, Inc., D. estimated that there were 191,000 exposed employees, 30,000 of whom exposed were ppm to 1 more of benzene. 1 Impact Economic Statement, p. 3-5, 11 Record, Ex. 5A, p. 3-5. explanation its permanent standard OSHA stated 1,440 there were exposed employees who worked in benzene plants, 98,000 petroleum in other refineries, 24,000 in ovens, coke 4,000 light plants, 2,760 oil petrochemical industry, 52,345 who worked in bulk terminals, 23,471 drivers who loaded benzene from terminals, 74,000 those gas oil and production, 17,000 in pipeline work, 100 at tank-car facilities, 200 at tank-truck facilities, 480 on barges, 11,400 in tire-manufacturing plants, 13,050 types other production. rubber 43 Reg. Fed. (1978). 5936-5938 Although gave OSHA no estimate for laboratory workers, the A. D. study Little indicated that there 25,000 were exposed workers in that industry. figures These up add 347,246 exposed em- ployees approximately 282,000 less than the overall — estimate 629,000. possible It is that some or all employees of these work in the “other indus- briefly tries” described in OSHA’s explanation; these are primarily small firms that adhesives, manufacture paint and ink or that use benzene sol- Id., at vents 5939. No estimate of the number of exposed employees in those industries or the aggregate of compliance cost by those industries D, given by either by OSHA or A Little its consulting report.

617 immediate an almost produces concentrations high Exposure concen- Inhalation of system. nervous central effect on the minutes; exposures fatal within 20,000 ppm can trations of nausea, and can cause ppm vertigo, to 500 range of 250 Reg. (1978). 5921 43 poisoning. Fed. of mild symptoms other may lead ppm 25-40 above exposures levels Persistent blood-forming organs, the and diseases of blood deficiencies generally fatal. which including aplastic anemia, ex- been aware that experts long have health Industrial nonmalignant may types various lead to to benzene posure connecting high levels the evidence By 1948 diseases. strong that become so had serious blood disorders benzene to lim- ppm a 35 imposed of Massachusetts Commonwealth the the In 1969 jurisdiction. within its workplaces itation adopted a (ANSI) Institute National Standards American averaged 8- over an ppm of 10 consensus standard national for 10- ppm of 25 concentration ceiling with a period hour ppm. of 50 peak concentration or maximum periods minute Id., Safety and Occupational the after at 5919. this consensus Secretary adopted the passed, Act was Health to 29 S. C. standard, pursuant U. the federal standard as (a).7 655§ 7 (a), provides: 655 Act, in 29 U. S. C. (a) as set forth Section of this subsections 5 of or to other regard chapter Title “Without period during practicable Secretary shall, as soon section, years ending after chapter two beginning date of this effective safety standard occupational or health date, by promulgate as an rule such standard, any Federal standard, established

any national consensus promulgation of such a standard he unless determines designated employees. specifically improved safety or health result standards, shall among such conflict In the event of safety greatest protection of assures promulgate the standard which employees.” of the affected or health most Secretary complied directive to choose with the In this case ppm, rather standard by selecting the ANSI protective standard than Gov- adopted the American Conference ppm Reg. Hygienists. Fed. Industrial ernment 618 As early as 1928, some health experts theorized that there

might also be a connection between benzene in the workplace and leukemia.8 In the late 1960’s and early 1970’s a num- ber of epidemiological studies were published indicating that workers exposed to high concentrations of benzene were sub- ject to a significantly increased risk of leukemia.9 In a 1974 report recommending permanent for benzene, the National Institute for Occupational Safety and Health 8See Delore & Borgomano, aigué Leucémie au cours de l’intoxication benzenique. l’origine Sur toxique de certaines leucémies aigues et leurs relations anémies graves, 9 avec.Ies Journal de Medécine Lyon de 227 (1928). A translation of that document appears in the benzene adminis trative record. 2 Record, Ex. 2-60. See also Hunter, Chronic Exposure (Benzol). Benzene II. The Clinical Effects, 21 Hyg. J. Ind. & Toxicol. 331 (1939), Record, 2-74, Ex. which refers to “leucemia” as a side effect of exposure chronic to benzene. 9 Dr. Muzaffer Aksoy, a Turkish physician who testified at the hearing on proposed standard, benzene did a number of concerning studies the effects of benzene on Turkish shoemakers. The workers in Aksoy’s Dr. studies used solvents containing large percentages of benzene constantly were exposed to high concentrations of vapors benzene (between 150 ppm) and 650 poorly under ventilated generally unhygienic Aksoy, conditions. See Acute Leukemia Due to Chronic Ex posure Benzene, 52 Am. J. of (1972), Medicine 160 1 Record, 2-29; Ex. Aksoy, (Benzol): Benzene Toxicity Its Effects the Hematopoietic System, Faculty Istanbul of Medicine Monograph Series (1970), No. 51 2 Record, 2-55; Aksoy, Ex. Erdem, DinCol, & Leukemia in Shoe-Workers Exposed Chronically Benzene, (1974), Blood 837 2 Record, 2-53 Ex. (reporting on 26 shoeworkers who had contracted leukemia 1973; this represented an incidence 100,000 of 13 per rather than the 100,000 cases per that would normally expected). Dr. Vigliani Enrico reported also an excess number of leukemia cases among exposed Italian shoemakers glues containing high percentage of benzene and in rotogravure workers plants exposed who had been over long periods of time to inks containing and solvents as much ben- 60% Vigliani Saita, zene. See & Leukemia, Benzene 271 New Eng. J. of Medicine (1964), Record, 2-27; 872-876 Ex. Vigliani, Forni & Chemical Leukemogenesis Man, 211 (1974), Ser. Haemat. 2 Record, Ex. 2-50.

619 studies these arm,10 noted that OSHA’s research (NIOSH), leuke- caused that benzene possibility” “distinct raised the cases had known all that of light in the fact But, mia. to declined NIOSH levels, exposure at very high occurred which con- standard, it ppm 10 change recommend diseases. nonmalignant against protect to sufficient sidered necessary to were studies that further NIOSH suggested between ben- link awas conclusively there whether determine were levels exposure so, if what and, leukemia zene and dangerous.11 published were 1976 studies additional 1974 and

Between can cause benzene that view confirm the to tended which In an high.12 are levels exposure least when leukemia, at Health, 10 Department (a) (3) requires the 669 29 S. C. Title U. § Department Health (now part in (HEW) Education, Welfare materials dealing with toxic Services) develop “criteria” Human are safe “exposure levels agents that describe physical harmful this sec obligations under HEW’s employment.” periods of for various 671. NIOSH, S. C. 29 U. delegated to have been tion 11 Standard— for a Recommended HEW, NIOSH, Criteria Dept, of See 1974), 1 74-137, (Pub. No. 74-75 Benzene Exposure Occupational the Office Director from the to a letter response In Record, 2-3. Ex. de ppm standard 10 its Division, stated NIOSH of Standards NIOSH risks. health leukemia, as other well against protect signed establish necessary order in research was however, further noted, Record, Ex. 12 and leukemia. for benzene dose-response data adequate 32A, 32B. additional reporting study 1976 in 12 Aksoy another published article, noted he also In that 1973. after cases uncovered eight leukemia a decline had led solvent as a benzene ban on use 1969 Aksoy, 1974. in beginning cases reported leukemia number in Haematologica Poisoning, Acta 55 Benzene in Chronic of Leukemia Types leukemia in a decline Vigliani also noted 2-30. Record, Ex. (1976), 1 65 See inks. glues and longer used was no Italy benzene after cases 122 Res. Leukemia, Environmental Forni, Benzene & Vigliani Benzene Associated 2-15; Vigliani, Leukemia Record, Ex. (1976), 1 Acad, Record, Ex. (1976), Sciences Y. Annals N. Exposure, pure ben- past Vigliani study noted the latter 100% 2-49. August 1976 revision of its earlier recommendation, NIOSH stated that provided these studies proof “conclusive” of a causal connection between benzene and leukemia. Record, 2-5, p. Ex. Although 100. it acknowledged that none of the intervening studies provided had the dose-response data id., had found lacking years earlier, two at NIOSH never- theless recommended that limit be set as low as possible. As a result of this recommendation, con- OSHA tracted with a consulting firm to do study the costs to industry of complying with ppm the 10 standard then in effect or, alternatively, with whatever standard would be the lowest feasible. Tr. 505-506.

In October 1976, NIOSH sent another memorandum to OSHA, seeking acceleration of the rulemaking process and “strongly” recommending the issuance of an emergency tem porary pursuant standard (c) 6§ of the Act, 29 U. S. C. (c),13 § 655 for benzene and two other chemicals believed to zene solvents had been used and workers exposed had been prolonged on a basis to concentrations ppm, of 200-500 peaks with up ppm. of- to 1500

A epidemiological number of were among studies also done American rubber during period. workers Dr. A. J. McMichael’s studies indi- cated a ninefold increase in the contracting risk of among leukemia work- heavily ers exposed who were pure the 1940’s and 1950’s to benzene used McMichael, as a Spirtas, solvent. Kupper, Gamble, Exposure & Solvent Among and Leukemia Epidemiologic Study, Rubber An Workers: 17 J. Occup. 234, (1975), Record, Andjel- Med. 2-37. See Ex. also kovic, Taulbee, Symons, Mortality Experience & of a Cohort of Rubber Workers, 1964-1973, Occup. (1976), 2 Record, 18 J. of Med. 387 Ex. (also indicating mortality among 2-54 an excess rate from leukemia rubber workers). (c) provides: Section 655

“(1) requirements provide, regard shall without to the 5, emergency temporary chapter title for an standard to take imme- 5 of Register upon publication in the Federal if diate effect he determines exposed grave danger (A) employees are to sub- physically be toxic or agents determined to harmful or from stances or 1 ppm exposure recommended that a carcinogens. NIOSH Record, Ex. Appar 2-6. imposed for benzene.14 limit be recommendation, OSHA asked of the NIOSH ently because complying cost of its consultant determine stand of with the “minimum feasible” standard instead ppm for voluntary guidelines It also issued Tr. 506-507. ard. 1to exposure levels be limited recommending that benzene, pos wherever time-weighted basis average 8-hour on an ppm 2 Record, Ex. 2-44. sible. two Pliofilm NIOSH had selected 1976, spring

In the epidemiological Akron, Ohio, Marys plants St. exposure. benzene leukemia and link study of the between report OSHA an interim forwarded April NIOSH expected inci- increase indicating at least a fivefold exposed to ben- had been workers who of leukemia for dence necessary pro- (B) emergency standard hazards, that such new danger. employees from such tect by a superseded effective until “(2) standard shall Such paragraph procedures prescribed with the in accordance promulgated (3) subsection. of this Register the in Federal standard the “(3) Upon of such publication (b) subsection proceeding in accordance Secretary shall commence pro- as a also serve published shall the as section, and standard this of a standard promulgate The shall proceeding. posed for the rule publication of the after than six months paragraph later no under (2) this subsection.” paragraph of in provided as emergency standard 14 representatives NIOSH standard hearing permanent on At with the initially in connection ppm1 they selected that had testified proceeding vinyl chloride. proposed standard for aof issuance detect level approximately the lowest ppm1 was they had discovered monitoring instruments. relatively unsophisticated through the use able appropriate ppm1 was an they thought that respect benzene, also With of the require the elimination might standard because lower nor that are up ppm) (in places 0.5 some of benzene amounts small recom NIOSH’s Tr. 1142-1143. atmosphere. present in the mally feasibility, either tech any evaluation on based mendation Id., at ppm. 1 exposures eliminating above economic, all nological or 1156.

622 sub- plants report

zene at two from 1940 to 1949.15 erroneously suggested exposures in the mitted OSHA plants ppm two between zero and 15 dur- generally had been ing period in As a of this new evidence question.16 result

15Seven out of the 748 workers fatalities from leukemia were discovered surveyed. However, Infante, study, Dr. stated that who conducted techniques probably his statistical had underestimated the number Id., actually leukemia cases that had at The normal ex occurred. 747. pected Record, 2 population in incidence leukemia such a would be 1.4. 2-51, p. Ex. 6. 16 exposure respect The authors’ levels was based statement with report by indicating that, after a 1946 Industrial Ohio Commission installed, exposures equipment some new at the St. ventilation had been Marys limits, plant brought instances had been within “safe” most Id., ranging ppm. ad from zero to 10 to 15 at 3. As the authors later mitted, ppm. 100 Tr. 814-815. the level considered “safe” in 1946 was begun Moreover, only had one of the seven who died of leukemia workers working Marys at 1946. the others had worked at St. after Five of Id., surveyed. at plant, employed Akron which workers report Department A of Health indicated 2537-2538. the Ohio exposure plant ppm, with excursions levels at the Akron of well over 84A, 61-62. 1,000 ppm. Record, App. A, pp. up in some areas tc Ex. Marys intervening years, testimony by Surveys well as taken in the as St. standard, 3432-3437, hearing proposed indi employees at the on the Tr. relatively high exposures plants may cated that both of have had through the 1970’s. hearing Industry representatives argued this evidence indi- at they very high, exposure had been that the levels had been cated Post-Hearing past. See epidemiological studies conducted in the

other (OSHRC), pp. H-059 23- American Petroleum Institute No. Brief for witnesses, however, 217-33, sim- Record, pp. 23-37. NIOSH Ex. question years could not levels for ply stated that actual not study however, determined; they agree, their should did ppm. risk 10-15 Tr. increase in leukemia proof a fivefold taken as standard, agreed permanent OSHA explanation of the In its 814-815. dose-response relationship could be no witnesses that the NIOSH with study: from inferred exposures area hearing there were demonstrated that at the “Comments ppm], exceeding at times study levels period these [10-15 during this personal per no moni- parts million. Since hundreds reaching values

623 and the prodding Record, continued NIOSH, 2-7, Ex. standard, OSHA did emergency issue an May effective 21. 1977, reducing exposure the benzene ppm limit from 10 1to ppm, ceiling for exposures up to 10 minutes from 25 ppm to ppm, and eliminating the authority for peak con- centrations ppm. of 50 Reg. (1977). Fed. In its explanation accompanying emergency standard, OSHA stated that benzene had been shown to cause leukemia at exposures ppm below 25 that, in light of consultant’s its report, was feasible to the exposure reduce 1 ppm. limit to Id., 22517, at 22521. May

On 1977, Appeals the Court of for the Fifth Cir- cuit entered a temporary restraining preventing order emergency standard from taking Thereafter, effect. OSHA abandoned its emergency efforts to make the standard effec- tive and proposal instead issued a for a permanent standard patterned entirely almost after emergency the aborted stand- Id., ard. at 27452. published its giving proposed statement notice

permanent standard, did ask OSHA for comments as to whether or presented not benzene a significant health risk at exposures of 10 ppm Rather, less. it asked for comments 1 ppm whether was the exposure minimum feasible limit.17 Ibid. Deputy OSHA’s As Director of Health Stand- ards, Wrenn, Grover testified the hearing, this formulation toring available, any data are regarding conclusion the actual individual time-weighted average exposure speculative. Because of the lack of exposure data, definitive OSHA linking cannot derive conclusions any specific excess leukemia risk observed level.” 43 Fed. Reg. sought public OSHA also comment as to whether certain industries compliance, should be exempt compliance pro proposed whether labeling techniques adequate, cedures and were what the environmental consequences regulation economic be, and whether it was replace products feasible to benzene in solvents and other of which it con stituted more than 1%.

of the issue tó be considered the Agency was consistent with OSHA’s general policy with respect to carcinogens.18 Whenever a carcinogen is involved, OSHA will presume no safe level of exposure exists in the absence of proof clear establishing such a level and will accordingly set exposure limit at the lowest level feasible.19 The proposed 1 ppm ex-

18It became clear at the hearing that OSHA had not promulgated the proposed standard in response any new concern about the nonmalig nant effects of low-level exposure. benzene See Tr. 126-127: “Is it accurate to say that why the reason why pro- OSHA has the — posed to exposure reduce the limits in the standard below the current levels is perceived because risk of leukemia, and not any because of new evidence it'has received that the current standards inadequate are protect against acute or chronic benzene toxicity, other than leukemia? “MR. WRENN: I think I simply will part refer my statement you were referring to, in which says, it is however benzene’s leukemo- genicity which greatest is of concern to OSHA. certainly That issue central within the ETS [emergency temporary and the standard] proposed standard.” 19Mr. Wrenn testified: proposed “The requires employee exposure to benzene in air be reduced to part one per million, with a part per five million ceiling any allowable over period fifteen during minute eight shift, hour work prohibits eye prolonged skin contact liquid benzene. “This airborne limit is based on OSHA’s regulatory established policy, that in the absence of a level, demonstrated safe or a no effect level carcinogen, it will be assumed that exist, none agency will attempt limit employee exposure to the lowest level feasible.” Id., at 29-30. See also: “MR. Wrenn, WARREN: Mr. in promulgating the emergency tem-

porary, proposed permanent, benzene standards, OSHA heavily, relies quoting I your am testimony now, on regulatory policy that there is no safe level for carcinogens any exposed population, —for and the fact leukemia, leukemogen and a carcinogen, is a is that correct? “MR. I WRENN: believe that I slightly stated that differently my summary oral statement than it is stated in the statement itself. I said absence of a known or demonstrated safe level or no

posure limit in this case thus was established not on the basis of proven a ppm, hazard at but rather on the basis of judgment “OSHA’s best proposal time of the of feasibility compliance of with the proposed industries.” Tr. 30. Given policy, [a]ffected OSHA’s cancer it inwas fact irrelevant whether there was any evidence at of all a leukemia risk at ppm. The important point was there nowas evidence that there risk, some how- small, ever at that level. The fact that OSHA did not ask on comments whether there exposure was a safe level of for benzene was indicative its further view that a demon- of such safety stration absolute could not be made.20 simply hearings Public were on proposed held com- standard, mencing July on 19, 1977. The final on standard was issued February 10, (1979).21 1978. 29 CFR 1910.1028 its form, designed final the benzene protect standard is to workers from whatever are hazards associated with low-level benzene level, policy exists, effect our regulate assume to that none and to Id., accordingly.” at 48-49. prefer I couple “MR. WRENN: would to it state as I have on a already morning, occasions and in a the absence of demon- exposure, strated level pur- safe we will assume none exists for Id., pose policy.” regulatory at 50. question In answer to the of what demonstration would suffice to level,” establish a “safe Mr. stated: Wrenn distinction, however, “I to would like draw between what I have referred exists, speculation to that a level as demonstration safe arid or elaborate may make, agency history theories that one and I think that in its very likely regulatory policy, would, future its face evidence carcinogenic exist, demonstrating that a hazard or did in this does exist very particular circumstances, accept set of reluctant as the decisions, regulatory argument basis for its theoretical that a level safe Id., may, particular fact, exist for a substance.” at 51-52. representative put succinctly, stating who testified later A NIOSH more leukemia, cancer, . if benzene causes that “. . if leukemia is a then Id., really is almost moot.” at 1007. promulgated 27, to the standard was An amendment June 1978. Reg. See n. Fed. 27962. infra. exposures by requiring employers workplaces to monitor determine the level exposure, provide medical examina- tions when the level rises above 0.5 ppm, to institute whatever or other engineering controls necessary are to keep exposures at or 1 ppm. below

In the standard originally proposed by the em- OSHA, ployer’s duty monitor, keep records, provide medical examinations arose whenever present benzene was in a workplace covered by the rule.22 Because benzene is omni- present small quantities, NIOSH and the President’s Coun- cil on- Wage and Price Stability the use of recommended “action level” trigger monitoring and medical examina- requirements. tion 1030-1032; Tr. App. 121-133. OSHA accepted this recommendation, providing under the final standard that, if monitoring initial benzene con- discloses centrations below ppm averaged 0.5 over an 8-hour work day, no further action is required unless there is change company’s practices.23 If exposures are above the action 22Apart from its gasoline storage exclusion and distribution facilities (an exclusion rule, text, infra), retained in the final pro see at n. posed rule also coverage excluded from liquid work operations in which mixtures containing or less year benzene were After a used. 1% exclusion was to be operations narrowed to where benzene solutions 0.1% were used. The rationale for the exposures exclusion was that airborne liquids from such generally However, be within 1 ppm limit. testimony at hearing proposed on the rule indicated that there was no predictable “consistent relationship” liquid between benzene content in a resulting and the exposure. Therefore, airborne OSHA abandoned the idea percentage of a liquid exclusion for benzene in its final standard. 43 Reg. Fed. position later and, OSHA reconsidered its an amendment to the permanent standard, liquids, reinstated for setting exclusion the level 0.5%, to be reduced years, id., after to three at 27962. 0.1% exemption monitoring from the testing portions medical workplaces for exposure the standard with benzene ppm levels below 0.5 predicated finding regulation any workplaces not was of such was consultant, Little, feasible. OSHA’s Inc., Arthur D. concluded that ppm exposure assuming a feasible limit even there was no level, ppm exposure limit, employers but below the are exposure monitor required quarterly to levels on a basis and provide to semiannual medical examinations for their ex- level, concept of nor posed employees. Neither an action challenged in this specific OSHA, level selected proceeding. employees monitoring initial indicates that are

Whenever 1 ppm above subject to airborne concentrations benzene ceiling 5 ppm with averaged workday, over an 8-hour modify to required employers period, 15-minute are controls to reduce ex- practice plants work their institute Consistent OSHA’s permissible limits. within posures to respirators regulation not allow policy, the does general technologically feasible.24 are if modifications engineering used perform required category are also Employers in this above workplaces remain long their monthly monitoring so exposed medical examinations provide ppm, semiannual “regulated areas” restrict access signs in and workers, post conduct exceeded, limit exposure permissible where necessary. where training programs employee liquid limits on strict places also The standard level was way, assuming that action (or, put another level action stated, practical decision was, NIOSH witnesses zero). Rather, as it below 0.5 are exposures that, benzene where on a determination based permissible unlikely to rise above the ever they be will ppm, that, absence concerned ppm. NIOSH was also 1of level might analytical equipment sophisticated used level, employers who action simply because provide medical examinations required monitor and to 1030-1032, 1133-1134. air. Tr. the ambient benzene in presence of 24 that an states OSHA Indeed, explanation in its example, (for in engineering controls institute required to employer is insufficient, hoods) controls are even if those stalling new ventilation therefore be used respirators must compliance themselves, achieve to engineering preference (1978). OSHA’s Reg. Fed. well. as rarely prop used respirators are opinion on its is based modifications etc.) fitted, properly uncomfortable, often not are they (because are erly adequate protective measures. considered cannot be therefore benzene. As originally framed, the standard totally prohib- ited skin or any eye any contact with liquid containing benzene. Ultimately, after the standard was challenged, OSHA this prohibition modified by excluding liquids contain- less ing than benzene. After three years, that exclusion 0.5% will be narrowed liquids to less containing than benzene. 0.1% permanent standard is expressly inapplicable to storage, transportation, distribution, sale, or gasoline use of or other fuels subsequent to discharge from bulk terminals.25 This exception is particularly significant in fact light 795,000 that over gas station employees, exposed who are an average of 102,700 gallons of gasoline (containing up benzene) are annually, thus excluded from the protection 2% of the standard.26

As presently formulated, the benzene expen- standard is an way sive of providing some protection additional for a rela- tively small number employees. According OSHA’s figures, the standard require will capital in engi- investments neering controls of approximately $266 million, first-year (for monitoring, operating costs medical testing, employee training, respirators) of $187 million to $205 million and inapplicable It is also operations to work involving liquid 0.5% (0.1% benzene years), supra, after three see n. handling and to the of benzene in systems, sealed containers or except employers insofar as required provide are cautionary appropriate employee notices and training. 26Prior to the introduction of the concept, action-level A. Little D. compliance estimated that for the costs industry might service station high $4 as Moreover, billion. Tr. 508-509. D.A. Economic Little’s Impact Statement indicated that employees service station generally were very exposed to low levels of Impact Statement, benzene. Economic 5A, p. 4-21, Record, p. Still, explanation Ex. 4-21. accompany its ing permanent standard OSHA did regulation not rule out of entirely, industry stating process was in the of studying whether regulate exposures extent it gasoline and to what should in general. 43 (1978). Reg. Fed. 5943

629 recurring $34 annual of approximately costs million.27 43 Fed. Reg. figures 5934 The in explana- outlined OSHA’s compliance tion of the costs to various industries indicate 35,000 only employees gain would any benefit regulation in terms a reduction in exposure their to ben- (24,450) zene.28 Over two-thirds these workers are em- ployed in the industry. Compliance rubber-manufacturing industry in costs are estimated to be rather low no capital operating costs and initial expenses only estimated at $34 million ($1,390 per employee); recurring annual costs low, would also be rather totaling By $1 less than million. contrast, segment refining of the petroleum industry that produces required $24 benzene would be to incur million in capital $600,000 costs and in first-year operating expenses to provide protection per additional for 300 workers .($82,000 employee), petrochemical industry while the would be re- quired capital $20.9 to incur million in million $1 costs and expenses employees in operating initial for the benefit of 552 Id., ($39,675 per employee).29 at 5936-5938. assump recurring OSHA’s estimate of annual costs was based on projected

tion that the levels it had initial would be confirmed that, monitoring year, be engineering after the first controls would bringing exposures these ppm successful in most within the limit. Under examinations, respira circumstances, monitoring, the need for medical would, course, drastically tors reduced. 5,000 in employees plants, Three hundred of these work in benzene refineries, 4,000 petro petroleum light plants, in 552 in the other oil laboratories, 1,250 industry, transportation, in chemical 156 in benzene 11,400 tire-manufacturing plants, 13,050 in other rubber-manu in (5,000 16,216 facturing plants. workers OSHA also estimated that another 1,104 industry, 7,300 bulk refineries, petrochemical petroleum laboratories) terminals, transportation, 2,500 in benzene ppm receive a benefit exposed of benzene and thus would to 0.5 to Id., comprehensive at 5936-5938. medical examinations. in terms of more per employee two industries is attributable high cost in the latter The allowing requiring engineering than policy controls rather to OSHA’s exposures permissible to the limit. to reduce respirators to be used 630

Although OSHA did not quantify benefits each category of worker in terms exposure benzene, decreased it appears from the economic impact study done at OSHA’s direction that those benefits may relatively Thus, small. although the exposure current is 10 the actual ppm, limit exposures outlined in study that considerably are often lower. For period for example, petro- 1970-1975 the industry reported chemical 496 that, out a total of em- ployees exposed only to benzene, exposed 53 were to levels 1 only between and 5 ppm (all 7 plant) same were exposed to between 5 1 ppm. and 10 Impact Economic Statement, p. 4-6, 4-2, Record, Table p. 5A, 4-6, Ex. Table 4-2. See also (indicating sample Tables 4.3-4.8 id., exposure levels industries). various

II point litigation critical issue at this is whether Appeals the Court of refusing was correct to enforce the ppm exposure ground supported limit by appropriate findings.30

relatively per employee industry low estimated cost in the rubber based on assumption OSHA’s solvents and can be other adhesives capital substituted for those that contain benzene will costs required. therefore not be Appeals correctly The other issue before us is whether the Court of refused to enforce dermal contact ban. That issue is discussed in IV, Part infra. respondents challenged monitoring In the court below also requirements, testing arguing medical that certain should industries have industries, totally exempt that, from them and as to other been requirements reasonably Agency all had not demonstrated that were safety. They necessary argued health and ensure worker also permissible exposure through requirement that limit be OSHA’s met through respirators reasonably engineering rather than was not controls ppm necessary exposure Because it invalidated the 1 under the Act. issues, they occasion to deal with Fifth Circuit had no these limit, the Court. before this are not now must, course, limit ppm exposure discussion

Any limit.31 imposing Agency’s rationale begin with the pages of the fills of the standard explanation The written to discussion Much it is devoted printed appendix. effects evidence of the adverse voluminous This ppm. 10 well of concentration above levels benzene at for justification ample that demonstrates there discussion to benzene that exposure occupational regulating (or peak a of ppm25 ceiling a 10 ppm, limit of prior direct provide however, not, It does 50 was reasonable. ppm) be should limit conclusion Agency’s that for the support ppm. 1 ppm from 10 to reduced adverse effects record in the administrative The evidence at best. OSHA sketchy ppm10 exposure at benzene nonmalignant certain dispute” that “no there was that noted of red the level by a reduction disorders, evidenced blood from could result blood, platelets cells or white studies several stated ppm. It then 25-40 exposures of blood in normal relatively slight changes had indicated per- ppm 25 below exposures result could values to make attempt not did OSHA 10 ppm. below haps risk significant of how these studies based estimate ppm of 10 exposures be at disease nonmalignant lack of data because stated that Rather, it or less.32 blood exposures and low-level between linkage concerning the dose-response a construct to impossible it was abnormalities, must determination agency’s validity held, have often weAs making that reasons agency’s stated the basis judged on (“[A]n ad 80, 95 Corp., Chenery S. 318 U. SEC v. See determination. upon which grounds upheld unless cannot order ministrative action upon which its those were exercising powers its agency acted 397; FTC v. Inc., S. v. Texaco 417 U. FPC sustained”); can be 233, 249. Co., 405 U. S. Sperry Hutchinson & by benzene caused abnormalities noted, blood some itself OSHA As may effects, while others health any discernible may have exposure (1978). Reg. 5921 death. Fed. and even impairment significant lead conclude, curve at this time.33 OSHA did however, studies demonstrated that ppm exposure current limit was inadequate single ensure that no worker would suffer a nonmalignant blood disorder as a result of exposure. benzene Noting that is “customary” set permissible exposure limit applying a safety factor of 10-100 to the lowest level at which adverse effects had been observed, Agency stated the evidence supported the conclusion that the limit should point be set “substantially ppm” less than even if benzene’s leukemic effects were not considered. 43 Fed. Reg. 5924-5925 OSHA did not state, however, the nonmalignant effects of benzene justified a re- duction in permissible exposure limit 1 ppm.34

OSHA also noted some indicating studies an increase in chromosomal chronically exposed aberrations workers dose-response “A curve shows the relationship expo between different sure levels and the risk of cancer other [or associated with disease] exposure those Generally, exposure levels. higher to levels carries with it higher risk, a exposure accompanied by lower levels is a reduced risk.” 2d, 504, 581 F. at n. 24. OSHA’s respect comments with insufficiency to the of the data were primarily addressed exposure the lack of data at low levels. OSHA did not discuss possible whether it rough was estimate, to make a based complete the more epidemiological and higher animal studies done at exposure levels, significance of the levels, of the risks attributable to those nor did it discuss possible whether it extrapolate was from such esti- mates to derive a risk exposures. estimate for low-level 34OSHA did not reducing invoke the automatic rule exposures to the lowest applies limit feasible that Instead, to cancer risks. prudent policy merely required reasoned that health permissible exposure “sufficiently limit be set below the levels at which adverse adequate effects have protection exposed been to assure observed for all employees.” Reg. 43 Fed. While OSHA concluded that application of this rule exposure “substantially would lead to an limit less ppm,” than 10 it did exposure not state either what level it considered to present significant a safety risk of harm or what applied factor should be permissible exposure level to establish a limit. ppm.”35 than 25 less “probably of benzene concentrations as to what position no definitive took Agency However, health of demonstrable meant terms aberrations these relation- dose-response quantitative no and stated effects circumstances, these Under established. yet been ship had as an “ad- categorized OSHA were effects chromosomal may pose which serious concern biological event verse considered must be such, as risk health potential reflect associated health effects purview adverse larger Id., 5932-5934. at benzene. e., (i. risk increased of an evidence leukemia, respect to

With due population) general by the borne than that greater risk sketchier. even was ppm10 or below exposures to benzene had relied study it NIOSH acknowledged that Once OSHA support not did emergency promulgating upon leuke- cause shown had been benzene view that its earlier supra, there see n. ppm, below at concentrations mia of such any evidence provided study that only one Chemical Dow by the study, conducted That risk. increased expected 0.2 versus deaths, leukemia three uncovered Co., appeared workers; it of a population out deaths, 2 to than more exposed to been never had three workers con- however, study, authors benzene. ppm relationship of a proof be viewed it could cluded all because leukemia benzene low-level between ato exposed occupationally been probably had workers three at other chemicals carcinogenic potentially other number had deaths leukemia no because careers in their points much exposed been who had among workers uncovered been permanent of explanation In its benzene. of levels higher three these that possibility that stated OSHA standard, not could exposure benzene by caused had been leukemias studies other noted that also studies, OSHA citing these While of chro level any increased indicated not had workers similarly exposed damage. mosome

ruled out and that the study, although not evidence of an increased risk of leukemia at 10 was ppm, therefore “consist- ent with the findings many studies that there is an excess leukemia risk among exposed benzene employees.” 43 Fed. Reg. 5928 (1978). The Agency no finding made that Dow study, any empirical other evidence, any opinion tes- timony demonstrated exposure that to benzene or below at ppm the 10 level had ever in fact caused leukemia. See 2d, F. at 503, where the Court of Appeals noted that OSHA was “unable point any empirical evidence documenting a leukemia risk at 10 . ppm. . .”

In the end OSHA’s rationale for lowering permissible exposure limit 1to ppm was based, any finding that not leukemia has ever been by caused to 10 ppm of ben exposure zene and that will it not be caused by exposure 1to ppm, but rather on series of assumptions indicating some leuke- mias might result from exposure ppm to 10 and that the num ber might cases reduced reducing the exposure level 1 ppm. In reaching result, the Agency unequivocally first concluded that benzene is a human carcinogen.36 Second, concluded industry had prove failed to there is safe threshold level exposure to benzene which no below excess leukemia cases would occur. In reaching this conclu sion rejected OSHA industry epide contentions that certain miological studies indicating no excess risk of among leukemia workers exposed at levels below 10 ppm were sufficient to estab lish that the threshold level safe or above 36“The conclusively evidence in the record establishes that benzene is a carcinogen. human The determination leukemogenicity benzene’s is de- rived from totality evaluation of all evidence in based on particular study. one recognizes, OSHA as indicated in- above reports vary considerably quality, dividual investigations that some significant methodological have deficiencies. recognizing the strengths While studies, and weaknesses in individual OSHA nevertheless concludes that clearly the benzene record as a whole relationship establishes a causal Id., between benzene leukemia.” at 5931. testimony that witness’ industry rejected an It ppm.37 also of the basis on constructed the curve could dose-response indicated that this curve studies epidemiological reported 1 ppm limit 10 to exposure the permissible reducing cancer other and one one leukemia at prevent most years.38 every six death respect policy Agency applied its

Third, the of definitive that, in absence concluding the carcinogens,39 epi “Although the that: studies, stated OSHA rejecting these relationship of a strong causal provide evidence can demiological method by its findings, it is positive of the case in exposure and disease between noting a After measure.” an insensitive very relativelytcrude nature defective, the often are studies ways which such specific in of number studies, negative evaluating policy when is “OSHA’s it Agency stated Id., accuracy." at methodological of higher standards hold them light, concluded OSHA industry in this studies Viewing the 5931-5932. un them make methodological defects sufficient had of them that each benzene. exposures to safety of low-level reliable indicators exposure it believed because testimony part rejected this OSHA a dose- inadequate to formulate studies to epidemiological data in accepted— testimony was if the even also indicated It response curve. Agency the risk of cancer —the any increase long as there indeed is most which exposure level “select the obligation to was under Id., at 5941. employees.” exposed protective Agency that the states Marshall Me. Justice dissenting opinion, In his ” setting a carcinogen ‘policy’ Draconian “blindly some rely on not did large number points to He for benzene. exposure limit permissible compiled as it record the voluminous Agency heard witnesses concerning benzene. facts particular on instead it relied evidence that interpre MaRshall’s Justice disagree with Me. respect, we all due With evi hearing the After its decision. for Agency’s rationale tation at stated it had policy view same dence, Agency relied of clear absence that, namely, 623-625, supra, outset, see for level exists that no safe be assumed contrary, must it to the evidence entirely predicta reached also Agency carcinogen. ato impossible concededly its carried industry had ble conclusion exposure exists level a safe infra, proving burden, see n. cancer generic proposed in its later clear Agency made theAs benzene. go industry witnesses to allow 51, infra, compelled felt see n. policy, despite carcinogen, dealing with regulation in each ground same over proof of a level, safe it must be assumed any that level above presents zero some increased risk of cancer.40 As the federal parties point out in brief, their there are a number of scien- public tists and specialists health who view, subscribe to this theorizing a susceptible that person may contract cancer from absorption of even one molecule a ben- carcinogen like zene. Brief for Federal Parties 18-19.41 policy its generic view. policy, yet effect, gone which has not into specifically was designed duplication to eliminate this of effort in each case by foreclosing industry arguing that there is a safe level for the particular carcinogen being regulated. (1977). Reg. 42 Fed. 54154r-54155 40“As above, stated positive studies on benzene demonstrate the relationship causal Although benzene to the induction of leukemia. studies, these part high the most exposure levels, involve it is OSHA’s view carcinogenicity that once the of a substance been established .has qualitatively, any exposure by must be considered be to attended risk considering when population. given OSHA therefore that believes occupational exposure poses to at carcinogenic benzene low levels a risk Reg. (1978). workers.” 43 Fed. 41The theory so-called “one hit” laboratory is based on studies indicat ing that carcinogen may one molecule a react in the test tube with one produce molecule of DNA theory that, a mutation. The is if this body, occurred in the human replicate the mutated molecule could over a period years eventually develop into a cancerous tumor. See Proposed OSHA’s Identification, Rule on the Regulation Classification and Posing Toxic Carcinogenic Substances Risk, Potential Reg. 42 Fed. Industry 54165-54167 challenged witnesses theory, this ar guing presence of several different defense mechanisms in the body unlikely human person make it actually that a contract cancer absorbing carcinogenic as a result of one Thus, molecule. the molecule might reaching site, be detoxified before damage critical to a DNA might repaired, molecule might or a mutated DNA molecule be de stroyed by body’s immunological defenses before it develop could into a cancer. Tr. 2836. light person’s contracting improbability of a cancer as a result single testifying hit,

of a a number of the on both scientists sides of the every probably agreed individual have expo- issue does a threshold g., id., he or she See, sure limit below which will not contract cancer. e. at problem, however, susceptibility 1179-1181. The individual ap- pears vary greatly present way and there is at no to calculate each stating Act, its view of the

Fourth, Agency reiterated either (b) (5) set the standard required safe been demonstrated level that has at the level If no safe higher. feasible, whichever is lowest level Secretary’s interpretation case, as in established, of an selection automatically leads to the of the statute of ben- Because feasible.42 lowest is the exposure limit suggested ever no one has economy, to the importance zene’s to entirely, or use to its eliminate would be that it feasible omni- that are small amounts to the exposures try to limit workable 1 as a ppm selected Agency Rather, the present. that determined then supra, and 14, see n. exposure level, feasible technologically level was with that compliance not be will [compliance] . of . impact . economic “the that firms the affected of welfare financial threaten the as to such It (1978). 5939 Reg. economy.” 43 Fed. general or the expo- feasible minimum 1 ppm was the that therefore held Act. of the (b)(5) § 6 of meaning the level within sure in its discussion refer did not Agency although the Finally, the identify to any duty authority to legal pertinent of the that conclude did standard, it the new benefits anticipated exposure reducing the from likely to result were benefits some based, This conclusion 1 ppm. 10 to ppm from limit the assumption that on the but rather evidence, again, levels decrease. exposure decrease will leukemia risk of a construct impossible it had found Agency Although accuracy any predict would that curve dose-response agreed that industry witnesses Thus, even person’s threshold. every every single certainty that with absolute ensure must if standard exposure limit only zero leukemia, any risk protected from worker is Id., 2830. at suffice. must, for the carcinogen is a benzene that doubt is no “There in Given such. workers, regulated as safety protection appropriate level, it is a safe or establish threshold ability demonstrate reduced exposure to benzene permissible prescribe OSHA (1978). Reg. 5932 43 Fed. feasible.” level lowest number of leukemias that could be expected to result exposures at 10 ppm, 1 ppm, at or any at level, intermediate it nevertheless “determined that the benefits of proposed likely standard are to be appreciable.” Beg. 43 Fed. In light of the Agency’s any ability disavowal of determine the numbers of employees likely adversely to be affected exposures of 10 ppm, Appeals held Court this finding to be unsupported by the record. 2d, F. 503.44

It is noteworthy point that at no lengthy its explanation did Agency quote or even (8) cite 3 made Act. It no finding any provisions of the new were “reasonably necessary appropriate provide health- safe or ful employment and places of employment.” Nor did allude to the possibility such finding might been have appropriate. point At an earlier explanation, its OSHA stated: general “There agreement benzene causes leukemia as well as other fatal of the bloodforming organs. diseases spite certainty of conclusion, adequate there does not exist an scientific establishing quantitative basis for the response dose relationship between exposure to benzene and the of induction leukemia and other blood uncertainty diseases. The magnitude in both the expected actual of deaths theory and in the of extrapolation from existing expo- data to the OSHA places sure levels the estimation of benefits on 'the frontiers scientific of

knowledge.’ While the actual estimation of the number of cancers to be prevented highly uncertain, is the evidence indicates that the number may appreciable. be general There agreement is that even in absence the ability to establish a ‘threshold’ or ‘safe’ level for benzene and other carcinogens, response a relationship likely dose exist; is, is expo- to that higher sure to higher doses with it cancer, carries a risk of and con- versely, exposure accompanied by to lower is levels risk, a reduced even though precise quantitative Id., relationship cannot established.” at 5940. did, however, The Agency’s court hold that the other conclusions— some that there is risk of at 10 ppm leukemia and that the risk would by decreasing decrease ppm the limit to 1 supported —were 2d, substantial evidence. 581 F. at 503.

63& Ill turns, large to a cases in these the issues of Our resolution relationship § between the meaning of and extent, on the standard as a safety standard and health which defines (8), provide safe to appropriate necessary and “reasonably is that the which directs (b)(5), 6§ employment/’ or healthful safety standard a health and promulgating Secretary in adequately which most the to “set materials toxic the best avail- basis on the feasible, extent to assures, the impair- material suffer will employee no that evidence, able . . .” capacity. functional or health ment term the definition (8)’s view, § Government’s merely requires best or at significance legal no has “standard” position takes It totally irrational. not a standard that to requires OSHA it that controlling (5) (b)6 is that assurance absolute gives either that a standard promulgate exposures reduces or worker every for each safety “feasi- interprets Government The feasible. level lowest a cost achievable technologically meaning ble” as subject to industries viability of the impair the not on the representatives, industry respondent regulation. correct Appeals Court argue hand, other appropriate” necessary and “reasonably holding requirement feasibility along with (8), 3§of language costs both quantify Agency requires (5), (b)6§ they are conclude rule and proposed of a benefits commensurate. roughly either whether necessary decide is view, In our we think it For entirely correct. industry Government standards permanent all apply (8) does 3§ clear Secre- the requires it Act under promulgated is it standard, to determine issuing before tary, significant remedy a to necessary appropriate reasonably Secretary Only after impairment. health of material risk exists risk such determination threshold made has with respect to a toxic substance, would it be necessary to de- cide whether § 6 (b)(5) requires him to select pro- most tective standard he can consistent with economic and tech- nological feasibility, whether, as respondents argue, benefits of the regulation must be commensurate with the costs its implementation. Because the Secretary did not make the required threshold finding in these cases, we have no occasion to determine whether costs must weighed against benefits in an appropriate case

A Under the Government’s view, §3 (8), if it has any sub- stantive content all,45 merely requires OSHA to issue stand- 45We cannot accept argument (8) totally is meaningless. § The Act authorizes Secretary promulgate three different kinds of standards —national standards, permanent consensus standards, tem porary emergency only standards. The substantive given criteria for two of these—national consensus permanent standards and safety standards for hazards not (b) (5) covered 6§ set forth in 3. —are While § is true 3 is entitled “definitions,” § that fact does not drain each definition of substantive content. For otherwise there would be no purpose in de fining the critical terms of the Moreover, statute. if definitions were ignored, there would statutory be no criteria at guide all to Secretary in promulgating either national consensus permanent standards or stand ards other than those dealing with toxic materials and physical harmful agents. mayWe expect Congress display perfect craftsmanship, but it is unrealistic to assume that it give intended to no direction whatso ever to Secretary in promulgating most of his standards. separate structure of the subsection describing emergency temporary standards, 29 (c), quoted U. S C. 655 supra, n. supports this con- clusion. It authorizes the bypass the normal procedures for setting permanent standards if he makes findings: (A) two that employees exposed are “grave danger” to toxic substances and (B) emergency that an “necessary” to protect the employees danger from that findings Those are to compared with those that are implicitly required by the definition permanent (A) that standard — there be a significant opposed to a “grave” risk, (B) —as that addi- — regulation tional “reasonably necessary or appropriate” opposed to —as “necessary.” It would be anomalous Congress require specific find- or more safer produce reasonably calculated to that are ards Apart Arg. 18, 20. of Oral Tr. environment. healthy work Government rationality, of requirement minimal from this power, Agency’s on the no limits (8) imposes 3§ argues that do to employers requiring fromit prevent not thus would and all eliminate necessary” to “reasonably be would whatever respect to With workplaces.46 from their any harm risks Government agents, the physical harmful and substances toxic (5)’s (b)6§on Relying position. more extreme an even takes assures... adequately most “which set a standard to direction health impairment material suffer will employee no that that contends Government capacity,” the functional guarantee that either standards impose to required Secretary is health material risk any free are workplaces possible as as close come or that small, however impairment, industries. ruining entire without doing so to completely eliminate to were statute of the purpose If the we would harm, of serious risk certainty absolute with interpret Secretary to for the proper would it agree clear it is we think But fashion. in this (5) 6 (b) (8)3§§ to employers require to designed was not the statute it is tech- whenever workplaces absolutely risk-free provide great not cost as the long so so, do to feasible nologically lan- both Rather, industry. entire destroy an enough history, legislative its as as well Act, structure guage as elimination, require intended it indicate harm. risks significant feasible, of far blanche carte Secretary a give but standards temporary for mgs standards. permanent requirement feasibility concede not does Government safety standards health (5) applies (b)6 of § sentence second However, even supra. n. See standards. substances toxic than other when “feasible,” term interpretation Government’s did, if require Agency allow still (8), would view its coupled entire long as cost, so great risks insignificant of even elimination jeopardized. viability industry’s B By empowering the Secretary to promulgate standards are “reasonably necessary or appropriate to provide safe or healthful employment places and of employment,” the Act implies that, before promulgating any standard, Secretary must make a finding that workplaces question are not safe. But “safe” is not equivalent of “risk-free.” There are many activities that we engage in every day —such driving a car or even breathing city air —that entail some risk of accident or material health impairment; few nevertheless, people would consider these activities “unsafe.” Similarly, workplace can hardly be considered “unsafe” unless it threat- ens the workers with a significant risk of-harm.

Therefore, before he can promulgate any permanent health or safety standard, the Secretary is required to make a thresh- old finding that a place of employment is unsafe —in the sense significant risks present are can or eliminated lessened by a change in practices. This requirement applies to permanent standards promulgated pursuant 6 (b)(5), § as well as to other types of permanent standards. For there is no reason why (8)’s § 3 definition of a standard should not be deemed incorporated by reference into § 6 (b)(5). The standards promulgated pursuant (b)(5) § 6 just are one species of genus governed standards by the basic requirement. section That repeatedly uses the term “standard” without suggesting any exception from, quali- fication general of, definition; on the contrary, it directs the Secretary to select “the standard” —that is to say, one possible various satisfy alternatives that the basic definition (8) in 3§ protective.47 most Moreover, requiring —that (b) (5) (a) Section parallels respect. this § (a) Section 6 re quires Secretary, faced with when a choice between two national con standards, sensus to choose protective the more 7, supra. standard, see n. (a)6 suggest Just as protective does not this more standard need not meet the of a definition national consensus standard set forth (9), in 3§ risk is significant finding of Secretary make a threshold granted power regulatory scope of the with the consistent pro- empowers (b)(5), which by §6 him agents physical for chemicals standards, not mulgate physical “toxic materials” for but generally, “harmful agents.” supported (b)(5) (8)3 interpretation §§ This 6 (g) example, § for Thus, Act. provisions the other in part that provides stand- establishing priority determining

“[i]n re- due give Secretary shall section, under ards safety mandatory need for the of urgency gard to the trades, industries, particular for standards health and stand- material toxic protective most the suggest that (b) (5) not 6 so does § in § aof "standard” definition to the conform not need ard promulgate Secretary to requiring the (b)(5), while rest of The §6 employee will no . that . . adequately assures that “most standard the contains also capacity,” or functional health of impairment material suffer degrees in differences Secretary consider should implying that the phrases risks. all of elimination simply a total than significance rather assures, adequately “most that one is to selected be Thus, standard the no evidence,” that available best the feasible, the basis the extent account into to take directed Secretary is also The will result. harm such may information such other and demonstrations, experiments, “research, attainment the addition consider “[i]n and to appropriate” the employee ... the for protection safety and health degree of highest standards, the feasibility of field, the data scientific latest available safety laws.” health other this under gained experience the make (8) of § our view that states Marshall Mr. Justice sentence first The disagree. We (b) (5) superfluous. in sentence § first standard protective highly a Secretary select (5) requires the (b)6 of § promulgated. should standard that has determined he once that the sense a standard such need is a there finding threshold threshold the unlike is not workplace the risk significant is a there Once harmful. agent is physical a toxic chemical finding (5) him (b) directs finding, § requisite threshold Secretary made has aof definition meets still standard protective the most to choose feasibility. consistent (8), under §3 standard crafts, occupations, businesses, workplaces or work environments.”

The Government expressly has acknowledged this section requires Secretary to undertake analysis some cost-benefit before promulgates any he requiring the elimination standard, of the most serious hazards first.49 If analysis such an must precede promulgation standard, seems manifest Congress intended, at a Secretary bare minimum, that the a significant risk find of harm and probability therefore a significant establishing benefits before a new standard. 6 (b)(8)

Section lends support analysis. additional to this requires That subsection when that, Secretary substantially alters existing an consensus standard, he must how explain the new rule will “better effectuate” purposes Act.50 If this was intended requirement meaning to be more than a less it must formality, impose Secretary be read upon duty to find that existing national consensus adequate protect continuing workers from a risk significant Thus, Secretary harm. case, the required to exposures find that permissible the current 49 “First, requires (g) 29 U. priori S. C. 655 to establish setting occupational safety ties in health and standards so that the more setting priorities serious hazards are In addressed first. such Secre the tary course, must, costs, of consider the relative and risks.” benefits Reply Brief argues for Federal Parties 13. The Government that the Secretary’s priorities setting subject judicial of under this section is not Arg. agree review. Tr. of Oral 23. While we that a court cannot tell the Secretary admittedly significant which of two regu risks he should act to section, first, along (8) (b) this (5), late with 3 and 6 that the indicates §§ Secretary’s power limit the requiring signifi Act does the elimination of cant risks. 50 (b) (8), (b) 6 as set in (8), Section forth 29 provides: S. C. 655 U. promulgated by a rule Secretary the substantially “Whenever differs existing national standard, Secretary shall, from consensus the at the time, publish Register in the why same Federal a statement of the reasons better adopted rule as will effectuate purposes chapter the the of this than standard.” the national consensus

645 of harm significant risk present ppm of 10 a level exposure workplace. the is unreason Act, it mandate of a clear

In absence give intended to Congress that assume able to would industry that American over power unprecedented (b)(5), 6 (8) 3§§ view Government’s from the result testimony a that Expert policy. cancer with OSHA’s coupled because carcinogen human is probably substance —either con have individuals because or animals has caused cancer jus exposures high extremely following cancer tracted —would risk of some poses the substance that tify conclusion mat no exposure minute how matter harm no serious as risk regarded the they that experts testified many ter how pervasive justify in turn would conclusion That insignificant. feasibility. constraint by the only limited regulation of substances thousands literally are there that of the fact light carcinogens identified have been workplace that used give theory would Government’s carcinogens, suspect produce might costs enormous impose power OSHA benefit.51 any, discernible little, if (1977), Reg. policy, Fed. generic cancer proposed OSHA’s pro its Under merely hypothetical. possibility indicates animal from proof quantum of certain there is posal, whenever —either a sub studies —that epidemiological from frequently, or, less experiments, temporary stand emergency level, an any exposure cancer causes stance provide employers requiring immediately, promulgated ard would low to the exposures to reduce examinations medical monitoring and along same issued then proposed rule A level. feasible est evi presenting effectively foreclosed objecting employers lines, with levels. current associated or no risk is little there dence CFR, Id., at 54154-54155 ; Part fact indicated regulation is proposed scope carcinogens, occupational potential 2,415 a list published has NIOSH Effects Toxic Registry A Subfile Carcinogens: Suspected NIOSH, 1976). OSHA 2d 77-149, ed. (HEW No. Pub. Substances of Chemical proved been have substances of these tentatively concluded has See regulation. subject full be should therefore carcinogens and 1978). (July Release, 78-625 USDL Press OSHA If the Government were correct in arguing that neither §3(8) nor §6 (b)(5) requires the risk from a toxic substance quantified sufficiently to enable to characterize it as significant in an understandable way, the statute would make such a “sweeping delegation of legis- power” lative might be unconstitutional under the Court's reasoning in A. L. A. Poultry Schechter Corp. v. States, United 295 U. 495, 539, S. and Panama Co. Refining Ryan, v. 293 U. S. 388. A construction of the statute that avoids this kind of open-ended grant should certainly be favored.

C legislative history also supports the conclusion that Congress was concerned, not with absolute but safety, the elimination of significant harm. The examples of indus- trial referred hazards to in the Committee hearings and de- bates all involved situations in which the risk unques- tionably significant. Tor example, the Senate Committee on Labor and Public Welfare noted that byssinosis, disabling lung disease caused by breathing cotton dust, affected as many as of the workers in carding or spinning rooms in 30% some American cotton mills and that many as 100,000 ac- tive or retired workers were then suffering from the disease. It also noted that statistics indicated 20,000 50,000 out of workers who performed had insulation work likely were to die of asbestosis, lung cancer, or mesothelyioma as a result of breathing asbestos fibers. Another example given of an oc- cupational health hazard that would be controlled by Act was betanaphthylamine, a “chemical so toxic any expo- sure at all likely to cause the development of bladder can- cer over a period years.” Eep. S. No. 91-1282, pp. 3-4 (1970); Legislative History of the Occupational Safety and Health Act (Committee of 1970 Print compiled for the Sen- ate Committee on Labor and Public Welfare), pp. 143-144 (1971) (hereafter Leg. Hist.).

Moreover, Congress specifically (b) amended (5) to make

647 require it not the perfectly clear does absolutely risk-free that would assure an promulgate standards (b) (5) the initial Committee bill workplace. Section provided that under this Secretary, in standards promulgating

“[t]he adequately which most subsection, shall set of the best available feasibly assures, the basis any impairment will suffer employee that no evidence, expect- life functional or diminished capacity, health or exposure to the regular has ancy employee if such even period standard for such dealt hazard 91st supplied.) (Emphasis life.” S. working his Leg. Hist. p. 39 242. Sess., (1970), 2d Cong., questioned Dominick Senate, Senator the floor of On stating: provision, of this wisdom up to that? to live ever going are we in the world “How where Florida a place going to do about we What are employee perish at getting are mosquitoes — But in Florida? mosquitoes may be thought there Are Wisconsin. flies Minnesota there are black by those bitten employees get if say that going we done have they will not been lives of their the rest we but do be, they will Probably all? harm at Hist. 345. (1970), Leg. Cong. Rec. 36522 know?” subsection.52 deleting the entire an amendment He offered then also made Dominick Senator bill, Committee criticizing the following observations: does, to establish apparently attempt, as this section “It is unrealistic impossibility and safety any hazards. Absolute utopia free Con- for the of this act in the only confusion administration create it will (1970), goals.” Cong. Rec. clearly unattainable gress to set 480. Leg. Hist. let us take language, thinking President, just about ask, I Mr. “But present time. a bus conductor conductor or streetcar who is a

fellow streets pollution we have process of the world, How in After sponsors discussions with the of the Committee bill, Senator Dominick revised his amendment. Instead of delet ing (b) (5) first sentence of 6 entirely, his new amend ment application limited the subsection toxic physical materials and harmful agents changed “any” and impairment of health to impairment.53 “material” In dis cussing this Senator Dominick change, noted that the Com mittee’s bill read as if a no that, standard had to “assure matter anybody what protect was the doing, him for rest of against any the his life foreseeable hazard.” Such an “unrealistic in standard,” stated, he had been sponsors tended the explained the he bill. Rather, the intention of bill, implemented by amendment, the was require to Secretary

“to use his best promulgate efforts to the best avail- standards, able in doing, and so he should take into ... anyone agents account that in working physical toxic process or during of the automobile accidents that we have all working day any trolley car, may be, driving one a bus or or whatever any can we set standards that will make sure he will not have risk to totally impossible put his life for the rest of his life? It is for this to be bill; yet Cong. Rec., in a it is in the committee bill.” 116 at Leg. Hist. 423. opponent legislation, may exag- As an Dominick have Senator (8)

gerated significance problem language since the already Secretary prevent trying was to sufficient “to establish utopia Nevertheless, Congress free from hazards.” the fact that allay the bill to Dominick’s concern demonstrates that amended Senator “clearly goals.” to it did not intend the statute achieve unattainable placement Dominick had also been concerned that the Senator require Secretary “feasibly” read “ban could be to all occu word health, injury, impaired some risk of pations which there remains way “adequately” “feasibly” to most expectancy,” since the assure life might prohibit entirely. occupation be to protection well absolute 366-367, amendment, 36530, Leg. Cong. Rec., Hist. his final he at by relocating feasibility problem requirement, attempted to cure this feasibly adequately most which assures” changing “the standard adequately assures, most extent which feasible.” “the standard agents might which may be harmful subjected, such conditions for the working life, rest of his so can that we something which get might at now, not be toxic if he works in it if time, a short but he works it the rest might very his life dangerous; and we want make things sure such are taken into consideration in establishing Cong. Rec., 37622-37623, standards.” Leg. Hist. 502-503.54 sponsors bill,

Senator one of the of the Committee Williams, agreed interpretation, with the and the amendment adopted. Domi reply parties argue

In their brief the that the federal re nick is not simply amendment means that argue harm; they quired insignificant to eliminate threats standards that (b) (5) requires Secretary 6 still to set § that risk subject any to employee ensure not even one will be that be.55 may risk how small that harm —no matter serious thought Congress not have argues that could Mr. Justice Marshall (b) (5), originally meaning as (8) any inasmuch as 6 had substantive § § toxic simply standards for drafted, and not to applied to all standards However, legislative as this physical materials and harmful substances. “toxic sub the words indicates, omission of history appears it that (b)6 original agents” draft physical from § and “harmful stances” noted, Committee entirely As Dominick (5) inadvertent. Senator was category only that limited apply to always that subsection to had intended these special section for Congress drafted a that The reason substances. thought it suggests, because not, Mr. Justice Marshall as substances Rather, was it areas. special protection in these need for was a that there regulating problems in special recognized that there were Congress because case, the risks are safety In the latter risks. opposed risks as health may former, not be obvious, the risks while generally immediate periods of time to exposed long for has been a worker until evident account that took was to ensure It particular substances. (b) (5). Congress enacted long-term risks § of these it is true that some 24-26. While Federal Parties Reply Brief for unimpor with the relative were concerned comments Dominick’s Senator supra, 647), quoted it is (see “fly” example at his injuries minor tance of interpretation This Congress’ at with express recogni- odds tion of futility trying workplaces totally to make all risk-free. Moreover, not even interpreta- OSHA follows this tion (b) of 6 (5) to its logical conclusion. if OSHA Thus, is correct only no-risk level for due leukemia exposure benzene (5) is zero if interpretation (b) its 6§of is correct, OSHA should have set the limit close as zero as But go feasible. OSHA did about not its task way. Rather, began it 1 ppm selected at level, least in part to ensure employers would not required to eliminate benzene concentrations greater that were little than the so-called “background” exposures experienced by the (cid:127) population large. at 14, supra. despite See n. Then, sugges- tions some labor unions that it was feasible at least some exposures industries ppm,56 reduce to well below 1 OSHA decided to apply the same limit largely all, matter of administrative Reg. 5947 convenience. Fed.

OSHA also deviated from its interpretation (b) own of 6§ (5) in adopting action level of ppm 0.5 below which moni- toring and medical In light examinations are required. of OSHA’s cancer policy, must have that some assumed employees would be risk exposures because of 0.5 below ppm. These employees presumably would thus benefit medical examinations, which might uncover benzene-re- problems. lated OSHA’s Agency consultant advised the technologically economically it was require feasible to provided. that such examinations Nevertheless, OSHA adopted level, largely an action insignificant because the ben- possibility major the clear that he was also concerned with remote of 52, supra. injuries, see n. 56 suggested ppm permissible exposure union a 0.5 One limit for oil (rather ppm ceiling time-weighted 1 than a average) refineries and a industries, level, with no use of an action exposure for all other Tr. id., ceiling ppm industries, a 1 limit for all Another wanted 1257. 3375-3376.

651 necessary- the performing and examinations giving of such efits cost.57 the substantial justify not monitoring did 1 with a beginning practicality to concessions OSHA’s im- concept level action an using limit and exposure ppm requiring statute as the of interpretation an adopt plicitly with entirely consistent is It risks.58 insignificant of regulation the requires also Act the to hold that interpretation this standard-setting area the endeavors to its limit Agency risks harm. significant eliminating to im- is history, it legislative the respect to with Finally, concern its expressed repeatedly Congress note that to portant over power much have too Secretary to allowing the about give Sec- to refused Congress Thus, industry. American unilaterally because plants down shut to power retary the Marshall, 445 Corp. v. Whirlpool see danger, imminent Secretary’s power circumscribed narrowly 1, and U. S. effort This emergency standards.59 temporary to issue evidence 57 record by the suggested also is level an action “A need animal naturally from occurs to benzene minimal some con occurring benzene Naturally 750-760). 749-750; (Tr. plant matter (Ex. per billion parts to 15 0.02 range may appears, centrations, it that their employers by certain suggested Additionally, it was 1). 117, p. because requirements exempted from operations exposures level low only intermittent involve operations these these accommodate concept should level action The use benzene. it extremely since low indeed are exposures where all cases concerns below are who employees monitoring of substantially reduces for medical requirements employees for these removes level action overexposure significant with employees time, the same At surveillance. added.) (Emphasis the standard.” protection full afforded are Reg. 5942 Fed. attempt policy OSHA’s also states Government certainly reasonable While possible. wherever benefits quantify duty under § its view own OSHA’s consistent it is not position, legis and the position in OSHA’s inconsistencies light of In (b)(5). interpretation. Agency’s defer Act, we decline history of the lative Labor, F. Dept. U. S. Assn., Inc. v. Peach Growers Florida intended Congress noted 1974), court (CA5 120, 130, n. 2d Congress to limit the Secretary’s power is not consistent a view that the mere possibility that some employee some- where in the country may confront some risk of cancer is a sufficient for basis of exercise the Secretary’s power to require expenditure of of hundreds millions dollars to minimize that risk.

D Given conclusion that the Act empowers Secretary to promulgate health and safety only standards where a significant risk harm exists, the critical issue becomes how to define and allocate the burden proving significance the risk in a case such as this, where scientific knowledge is imperfect and the precise quantification of risks is therefore impossible. The Agency’s position is that there substantial evidence in support the record to its conclusion no that there is absolutely safe level for a carcinogen that, therefore, the burden is properly on industry to prove, apparently beyond a shadow of a doubt, that there is a safe level for expo- benzene sure. Agency argues that, because of the uncertainties in this area, any other approach would forcing render it helpless, it to wait for the leukemia deaths that it likely believes are occur60 taking any before regulatory action. to restrict emergency standards, use of promulgated which are without any hearing. notice or that, It held promulgating emergency standard, only danger OSHA must find a or even some danger exposure, grave but danger also a exposure necessitating from Dry emergency Accord, Assn., Dept. Color action. Inc. v. U. S. Mfrs.

Labor, (CA3 1973) (an 2d emergency F. standard must be supported by something possibility more than may that a substance man). cause cancer in Congress carefully Secretary’s also powers circumscribed enforcement by creating new, independent appeals board to handle from citations noncompliance issued safety with health and stand- ards. See 29 S. C. 659-661. U. §§ above, acknowledged As noted OSHA empirical there was no evidence to the conclusion support there was risk whatsoever of exposures ppm. due deaths upon theory OSHA relied What was a

653 the on statute, burden was we read the the disagree. As We show, it evidence, that the basis substantial to on of Agency exposure to long-term that likely than not is at more least health of material risk significant a presents benzene ppm of a rule of proponent is the Ordinarily, it impairment. proceed- of in administrative proof burden who has the order toxic involving (d). In some cases S. ings. U. C. See a that proving of the burden shifted substances, Congress has pro- the opposing party safe onto substance particular course follow this not Congress did that The fact posed rule.61 Act indicates Safety and Health Occupational enacting the of burden normal bear Agency to it intended that proposed for a standard. the need establishing carry its burden attempt did not even OSHA In this case that finding benzene making a came to it The closest proof. its was workplace of harm risk significant a presented lowering the from derived be benefits to that statement "likely” were ppm1 10 to exposure level permissible held Appeals The Court “appreciable.” be Of evidence. substantial by supported not finding was this by substantial supported if were even importance, greater satisfy the sufficient be finding would evidence, such a Act. under the obligations Agency’s be perhaps can findings Agency’s inadequacy exposures, higher at much occurred deaths had that, leukemia because exposures. relatively low likely at fewer) to occur were also (although some number its conclusion held specifically Appeals The Court supra, See by record. unsupported was appreciable “likely” at 638. EPA, D. C. App. Fund, U. S. Inc. v. Environmental See Defense denied, 431 (1977), cert. 998, 1004, 1012-1018 2d 49, 57-63, 548 F. 43, has the- the EPA argument rejected the the court where U. S. regis suspend its in order unsafe pesticide proving that burden Act. and Rodenticide Insecticide, Fungicide, the Federal under tration ordinary bur deliberately shifted Congress had noted The court requiring manufac Act, Procedure Administrative under proof den products. safety of their the continued establish turers to illustrated best rejection its of industry testimony that a dose-response curve can formulated of current basis ' epidemiological that, evidence and even under the most con- extrapolation

servative theory, current levels would cause most two of population deaths out 30,000 of about every workers years. six See 38, supra. In rejecting n. testimony, OSHA made following statement:

“In the face of the record evidence of numerous actual deaths attributable to benzene-induced leukemia other fatal blood diseases, rely OSHA is on unwilling to hypothesis the that at every years most two cancers six prevented by would be proposed By way the standard. example, of study the Infante leu- disclosed seven excess kemia deaths population in a people over a about 600 25-year period. study higher While the Infante involved exposures then currently incidence those the encountered, by together rates found Infante, with the numerous other reported cases the literature of benzene leukemia and diseases, other fatal blood make it for to difficult OSHA rely on the hypothesis statutorily to assure the [witness’] protection mandated of employees. any event, due to the fact that there is no ben- exposure safe level of to impossible zene and it is precisely quantify to anticipated benefits, OSHA must select the level of ex- posure protective which is most exposed employees.” Reg. 43 Fed. possible

There three interpretations are stated of OSHA’s reason for rejecting testimony: (1) the witness’ con- OSHA probable greater sidered that a number of lives would saved lowering (2) from 10 ppm; standard OSHA thought that saving every years two lives six in a work force 30,000 persons is significant savings a reason- that makes it appropriate adopt standard; (3) able to a new if even is significant small number if savings not and even may smaller, Agency be even nevertheless believed it had is most the level duty select statutory to a economically if it employees exposed of the protective Secretary did if do so. Even to feasible technologically his construc- theory, third entirely on this rely to intend not so. do him to for proper make it statute tion sup- that would of fact findings express no Moreover, he made Under theory. drastic less ppm1 his port Govern- with the hardly agree can circumstances, we these Act. duty under its discharged OSHA ment a imposing contentions, the Government’s Contrary to risk significant demonstrating a Agency on the burden nor carcinogens, regulate ability its strip it of will harm before occur deaths for wait Agency require will it “significant” requirement First, any action. taking It straitjacket. mathematical is not identified risk instance, first determine, responsibility Agency’s are risks Some risk. “significant” abe it considers what If, unacceptable. plainly are others acceptable plainly will a person billion ain one are odds example, the water, the of chlorinated a drink taking by cancer die other theOn significant. considered be not could clearly risk inhalation regular ain thousand one are if the odds hand, fatal, a rea- bewill 2% benzene are vapors gasoline of significant risk the consider well might person sonable Although it. or eliminate decrease steps to appropriate take of probability exact calculate duty to no has Agency risk significant a find that obligation have does it harm, as employment of place a characterize can before present 62 “unsafe.” Marshall Justice 706, Mr. post, 62 at opinion, dissenting his acceptable determination involves question “[W]hen states: considera on necessarily be based must decision risk, ultimate level determina Factual facts. verifiable empirically well policy as tions judgment way; statistical some risk define most can tions resolution solely aon based be cannot tolerable risk whether Second, OSHA is required support its finding that a significant risk exists with anything approaching scientific cer tainty. Although the Agency’s findings must supported by substantial evidence, 29 U. S. §C. (f), (b)(5) specifi cally allows the Secretary to regulate on the basis the “best available evidence.” As several Courts of Appeals held, have provision requires a reviewing court give OSHA some leeway where its findings must be made on the frontiers of scientific knowledge. See Industrial Union Dept., AFL-CIO v. Hodgson, 331, S. App. U. D. C. 340, 499 2dF. 467, 476 (1974); Society the Plastics Industry, OSHA, Inc. v. F. 2d (CA2 1975), cert. denied, 421 U. S. 992. Thus, so long as they are supported by a body of reputable scientific thought, the Agency is free to use conservative assumptions in interpreting the data with respect to carcino gens, risking error on side of overprotection rather than underprotection. 63

Finally, the record in this case and OSHA’s own rulings on other carcinogens indicate there axe a number of ways in which the Agency can make a rational judgment about the facts.” agree. We Thus, while Agency must support finding its that a certain level of risk exists evidence, substantial recognize we its determination that a particular level of risk is "significant” will largely based policy considerations. At this point we have no need to reach the issue of what of scrutiny level reviewing apply court should to the type latter of determination. 63Mr. Justice Marshall that, states under our approach, Agency must either wait for deaths to occur or must “deceive the public” by mak ing basically meaningless significance determination of totally based on inadequate evidence. Mr. Justice view, however, Marshall’s rests on premise erroneous only reason OSHA did not attempt to quantify benefits in this case was because it could do not inso reason able manner. theAs discussion of Agency’s rejection industry of an attempt formulating at however, a dose-response curve demonstrates, see supra, 653-655, at Agency’s rejection of such methods dose-response curves part was based at in least on its view that nothing less than safety absolute would suffice.

657 significance relative with exposure risks associated to a particular carcinogen.64

It should noted in that, setting permissible also be a expo- sure in less-than-perfect level reliance on methods, OSHA would have the a backstop the form of monitor- benefit

64 example, For standard, in the coke-oven emissions OSHA had calcu 21,000 exposed lated that coke-oven workers had an annual excess mor tality of over 200 and proposed might that the standard well eliminate the entirely. Reg. 46742,46750 risk 41 (1976), upheld Fed. Iron American & OSHA, (CA3 granted, Steel 577 1978), post, p. Inst. v. F. 2d 825 cert. 909. hearings standard, Wage coke-oven emissions the Council on Stability year, Price estimated 8 35 lives would be saved each out 14,000 population workers, proposed of an estimated as a result of the Although noting range vary depend standard. benefits ing assumptions finding used, on the OSHA did not make a as to whether correct, ground its own staff estimate or CWPS’s was on the it was required quantify expected weigh or to benefits of the standard against projected those benefits costs. Agency proceedings, good

In other has had a deal of data from experiments significance animal on which it could on the base a conclusion example, vinyl of the risk. For the record on the chloride standard in significant developed a dicated that number of animals had tumors of liver, they vinyl lung, exposed ppm skin were chloride when to 50 during period 11 months. One died over a hundred out of animals Reg. (1974). Similarly, period. 35890, 35891 in a 39 Fed. regulating carcinogens, found that one of the substances OSHA lung ppm ppm, while had caused cancer in mice or rats at and even 0.1 subjected high the animals doses. another had caused tumors 80% Synthetic Organic Id., 3757, Chemical Assn. v. 3756, upheld in Mfrs. (CA3 denied, Brennan, 1974), 420 U. S. 2d 1155 cert. 503 F. (CA3 Synthetic Organic Brennan, Chemical Assn. F. 2d 385 v. Mfrs. denied, 1974), cert. 423 U. S. 830. studies, be benefit of animal Agency did not have the In this case yet experimental leukemia in have been unable as induce cause scientists did, however, have a fair exposure.- It result of benzene- animals as a negative evidence, including positive both epidemiological amount of Agency that this evidence was insufficient Although the stated studies. risks, and cancer between levels precise correlation construct determining likely whether it is more than helpful in at least be it would significant ppm. risk at there is not that 658 Thus,

ing testing. properly if OSHA medical determined exposure ppm, limit 5 it permissible that the should be set at for em- require monitoring testing could still and medical keep By could ployees exposed doing so, to lower levels.65 it assumptions made validity check on a constant sound exposure giving developing permissible limit, it initially set it evidentiary decreasing the limit if was basis workers way ensure that high.66 Moreover, too it could removed could be unusally susceptible who to benzene were any permanent they suffered exposure before had damage.67

E more de- involved a our of these cases has Because review must customary, it of the record than tailed examination 65 Comm’n, Safety Corp. Occupational Review v. and Health See GAF upheld 20, (1977), where the court App. 183 U. D. C. 561 F. 2d 913 S. medical required employers provide to the asbestos standard insofar as they fibers, if even employees exposed asbestos examinations for to permissible exposure limit. exposed were to concentrations below disputed OSHA’s industry representatives never respondent have although general, require monitoring power and medical examinations this case. specific requirements imposed in they object did to some of the case, we have no supra. disposition our n. Because of See objections what cost- or to determine pass specific these occasion to imposition of considerations, Agency’s any, govern the if should benefit requirements. such 66 that Con information-gathering function type precisely This is (b) empowers (7), the Secre 6 which gress mind when it enacted had in exposed employees require to be furnished tary medical examinations effectively de order to most potential hazards “in certain hazards adversely such employees is affected the health of such whether termine 91-1282, (1970), Leg. 147. p. Hist. Rep. No. See S. exposure.” was there OSHA noted the final standard explanation of In its disappear after testimony abnormalities that blood some however, refused Again, OSHA Reg. Fed. had ceased. requiring always inYet, occur. this would rely hypothesis on the ppm, ppm 0.5 employees exposed to between examinations medical backstop. the same kind of essentially itself with providing OSHA deter- any factual made neither have that we emphasized find- any factual own, rejected we have nor our of minations what on opinion express no Secretary. We the by made ings basis on the either might support, this record findings factual nor testimony; expert of basis or on the evidence empirical of of question difficult the more on any opinion express we do conclusion warrant determinations factual what of a promulgation make which present risks are significant appropriate. necessary or reasonably standard new actually findings the by supported course, be of must, believe we findings that merely Secretary, by the made made. have might he the clear perfectly makes record the case In this carcinogens for policy special aon squarely Secretary relied the existence proving industry of on burden the imposed Secretary’s the avoiding thereby exposure, level a safe more for need the establishing responsibility threshold author- statutory his interpreting so In standards. stringent power. his exceeded ity, the IV con- dermal the proceedings, administrative Throughout proposed In its attention. relatively little received issue tact contact eye skin on ban a total recommended OSHA rule dealing “in policy its the basis on benzene liquid e., (i. exposure routes potential all carcinogen, with a limited absorption) [should] skin ingestion, inhalation, There (1978). Reg. Fed. extent feasible.” hearing on requirement opposition little also rule proposed because apparently rule, proposed level permissible both from excluded mixtures involving liquid operations work ban contact dermal benzene. or less 0.1%) year, one after (and containing 1% per- eliminated OSHA however, standard, final its there ground benzene, liquid exclusion centage ben- percentage between correlation predictable nowas zene in a liquid and the airborne exposure arising from it. See 22, supra. n. Although the extent to which liquid ben- zene is absorbed through the skin is concededly unknown, OSHA also refused to exempt any liquids, no matter how lit- tle they benzene contained, -the ban on dermal contact. In support of this position it stated that there was no evi- dence to “suggest that the absorption rate depends on the amount of present benzene in the liquid.” Reg. 43 Fed. 5948-5949

After permanent standard was promulgated, OSHA received a number of requests from various industries that the percentage exclusion for liquids containing small amounts of benzene be reinstated. Those concerned with airborne exposures argued they should not be required to monitor workplaces simply they because petroleum-based handled products in which is an benzene unavoidable contaminant. Others concerned with the dermal contact ban made similar arguments. In particular, tire manufacturers argued that it impossible was for them to- comply with the ban because gloves cannot be worn during certain tire-building operations in which solvents are used and solvents containing absolutely no benzene are not commercially available.

Because of these requests, OSHA held a new series hear- ings and promulgated an amendment to the rule, reinstating the percentage exclusion, but lowering it from proposed The Agency did, however, provide 1% 0.5%. for a 3- year grace period before the exclusion dropped to 0.1%, rather than year the one that had originally been proposed. explaining its amendment, OSHA reiterated its policy with respect to carcinogens, stating that, because there is no abso- lutely safe for level any type of exposure, exposures by what- ever route must limited to the extent feasible. For air- exposures, borne a permissible zero exposure limit not had been feasible. However, in most industries a ban on any der- mal contact was feasible compliance since could be achieved simply by the of protective use clothing, such as impermeable ban the dermal contact

gloves. Agency recognized manufacturers, stated that a for but present problem could tire problem, be- would alleviate the exclusion percentage or less benzene were available containing solvents cause 0.5% solvents con- Although it noted that quantities. in sufficient in quan- available less benzene were not taining then 0.1% period would 3-year grace stated that a tity, Agency of solvents production to “allow time for increased be sufficient development lower amounts of benzene and containing with the compliance methods of and evaluation alternative Id., at 27968-27969. provision.” dermal standard’s con- the dermal down Appeals struck The Court it held that First, grounds. on two prohibition tact would result ban finding support record did not risk; leukemia reduced benefits terms quantifiable mean- necessary” within the “reasonably was not therefore, held that the court Second, of the Act. ing (8)3§of through absorbed may be benzene conclusion that Agency’s re- available evidence as on the best based the skin the second at 505-506. On (b)(5). 2d, 581 F. quired on the issue noted that the evidence the court ground, equivocal, through benzene the skin was absorption of some it be absorbed and indicating that could some studies All were rela- could not. of these studies indicating that the issue only expert who had testified on tively old determine, a now available to simple test was stated that what great accuracy, whether and to extent with a deal which re- light 16(b)(5), will absorption result. basis of promulgate standards on the Agency quires “the available scien- available evidence” and latest the “best held where there is field,” the court that tific data a will testimony simple test resolve uncontradicted information Agency required acquire issue, regulations require, which would “promulgating before change long-followed processes work industry established demonstrably 581 F. at 508. are not unsafe.” 2d, While the court below may have been in holding correct that, under peculiar circumstances of this case, OSHA was required to obtain more information, there is no need for us to reach that issue. For, order to justify a ban on dermal contact, the Agency must find that such ban a is “reasonably necessary and appropriate” to a significant remove risk of harm from such contact. The Agency did not such make a finding, but rather on acted of the basis absolute, no-risk policy that it applies to carcinogens. Indeed, on this issue the Agency’s position is even more untenable, inasmuch itas was required to only assume not that benzene in small doses ais carcinogen, but also can that it be absorbed through the skin in sufficient amounts present a carcinogenic risk. These assumptions are not a proper substitute for the findings a significant risk of harm required by the Act. judgment Appeals Court remanding peti-

tion for review to the Secretary for further proceedings is affirmed.

It is so ordered. Mb. Chief Justice Burger, concurring.

These cases press upon the Court difficult ques- unanswered tions on the frontiers of science and medicine. The statute and the legislative history give ambiguous signals as to how the Secretary is operate directed opinion this area. The takes a difficult task to decode Justice Mr. Stevens message guidelines statute as to for administrative action. comply

To with statutory requirements, must bear the burden “finding” proposed health safety standard “reasonably necessary or appropriate to provide safe employment places healthful of employ- ment.” policy judgment This subsidiary entails the finding the pre-existing presents “significant risk” of material health impairment for a spends worker who his entire employment in a working life environment where ex- The levels. Secre- permissible maximum at remains posure sufficiently “quantified finding of “risk” must factual tary’s significant as characterize it Secretary the to enable to Ante, Precisely what this 646. way.” understandable find- mandated because these say. But to difficult means is ppm I that the Secretary, agree the made were ings I However, would invalidated. must be benzene and the administra- of the courts functions differing the stress regulation. safety and such health respect with to agency tive func- the narrow regulator, and the ultimate Congress is The of statute and meaning the to discern courts is of the tion ensuring of objective with regulations implementing Secretary safety standards and health in promulgating fac- pertinent of each consideration given reasoned “has Permian statutory commands. with complied has and tors” hold- (1968). Our 747, 792 Cases, U. S. Area Rate Basin care greater steps his retrace Secretary must ing that the scope of the derogation not to be taken is consideration argu- facts and When discretion. agency legitimate of Secretary considered, the duly been presented ments have risk specific a whether toas judgment policy make must objec- policy in terms significant is impairment health pursuant capacity, he acts When the statute. tives he exercises by Congress, delegated authority legislative to the aspect only one focus legislature prerogatives —to that, regulations promulgate or to problem, larger of a alloca- inefficient or policy imprudent may appear some, not extend does function judicial resources. tion lies function That policy. regulatory revision substantive oversight Executive Congressional elsewhere —in boundaries be sure amendatory legislation although — indistinct. ill-defined often are stat- under duties his discharging when Nevertheless, remember admonished well ute, in this Inherent authority. his burdens heavy responsibility regulation refrain authority to statutory scheme 664

insignificant or de minimis risks. See Alabama Power v.Co. Costle, 88-89, S. App. 204 U. D. C. F. 636 2d 323, 360-361 (1979) (opinion Leventhal, J.). When the administrative record only reveals scant or minimal risk of material health impairment, responsible administration calls for avoidance extravagant, comprehensive regulation. safety Perfect is a chimera; regulation not must strangle human activity in the search for the impossible.

Mr. Justice Powell, part in concurring and concurring in judgment. I join Parts I, II, III-A, III-B, III-E III-C, and of the plurality opinion.1 Occupational The Safety and Health Ad- ministration relied in large part on its “carcinogen policy”— which had not been adopted formally promulgating —in exposure benzene and dermal regulation contact at issue these cases.2 For the reasons stated I plurality, agree (b)(5) §§ 6 3(8) of the Occupational Safety and Health Act of 1970, U. S. 29 C. (b)(5) §§ 655 (8), must be read together. They require OSHA to make a finding threshold proposed occupational health standards reasonably are necessary provide workplaces. safe When OSHA acts reduce existing national standards, consensus portions These plurality opinion primarily spe address OSHA’s carcinogen policy, cial than argument rather OSHA’s also made it evidentiary findings. I necessarily do every agree with observation plurality opinion concerning presence or absence of such find ings. express I also question no view whether a different inter pretation statute nondelegation would violate doctrine of L. Poultry Corp. A. Schechter A. States, v. United (1935), 295 U. S. Refining Ryan, Panama Co. post, v. 293 U. S. See (Rehnquist, J., 672-687 concurring judgment). promulgated of Labor pursuant the relevant statutory authority. his agency Since OSHA is the responsible for developing regulations such Secretary’s under direction, opinion agency” refers to “OSHA” or “the directly as the decisionmaker most concerned. (i)

therefore, currently permissible find that must impairment; health risk material significant create levels reduce significantly levels (ii) reduction those the hazard. nec- the that possibility the out rule I not

Although would properly policies generic on part could rest in findings essary Procedural McGarity, Substantive OSHA, see adopted Policy Science Resolution in Administrative Discretion OSHA, 67 EPA and Carcinogens Regulating Questions: agency supported no (1979), properly 729, 754-759 L. J. Geo. agree I therefore these cases.3 inus before policies are extent to the invalid is regulation the that plurality the carcino- known to exposure assumption the upon rests or, to be safe proved level to a reduced always should gens affected the level that the lowest found, to is if no such level technology. with available industry can achieve I exclusively based were regulation disputed If the need we agree that I also policy,” “carcinogen determine OSHA requires Act whether consider reasonably related are standard proposed of a benefits Ap- Court Ante, at 615. As compliance. costs takes the OSHA however, recognized, Fifth Circuit for the peals by specific justified regulation its position” “fall-back com- evidentiary record the voluminous findings upon based OSHA, Institute v. Petroleum American case. in this piled num- example, found, OSHA 493, 2d 503. 581 F. carcinogens effective regulating policy for adopted a formal has OSHA CFR, (to at (1980) be codified Reg. Fed. 1980. April promul agency in effect when policy was 1990). such But no Part determinations factual Moreover, neither regulation. its gated benzene sup rests are policy upon which judgments administrative nor occasion have no Accordingly, we record alone. adequately on this ported basis may supply a policies agency which valid the extent consider particular cases. exist health risks finding that for a ber of prevented cancers by reducing permissible exposure levels from 10 ppm to 1 ppm “may be appreciable,” “the benefits of proposed likely are apprecia- ble,” and that the “substantial costs the new [of standard] are justified in light of the hazards.” 43 Reg. Fed. 5940- Thus, OSHA found —at generally least —that the hazards of benzene currently permissible levels are serious enough to justify expenditure of hun- dreds millions of dollars. For me, that necessarily finding subsumes the conclusion that the health risk “significant.” If OSHA’s conclusion is supported by substantial evidence, the threshold requirement discussed in the plurality opinion would be satisfied.

IAs read its opinion, the does not plurality consider whether agency’s findings supported are by substantial evidence. The Court Appeals found them insufficient because OSHA failed “to estimate the extent of expected benefits. . . .” 581 F. 2d, at 504. That court apparently would required have OSHA to supply a specific numerical estimate of benefits through derived mathematical techniques quantifi- for “risk cation” or “cost-effectiveness analysis.” Id., at 504, 23; n. id., see at 504-505. I agree do not with Ap- the Court of peals’ conclusion that the statute requires quantification every risk in case. preference statutory for the “best available evidence,” U. S. C. 655 (b)(5), implies that OSHA use must

best known techniques for the accurate estimation of risks and benefits when such techniques are available. But neither the statute nor the legislative history suggests that OSHA’s hands are tied when quantification reasonable cannot be ac- complished by any known post, methods. See at (Mar- J., dissenting). In this litigation, OSHA found that shall, impossible “it to precisely quantify the anticipated bene- fits. . Reg. . .” Fed. If this finding sup- ported by substantial evidence, the statute does prevent not significant finding health hazard on the I testimony opinion. do weight expert basis ante, otherwise. See plurality to hold understand the 662. at I that “OSHA not hold foregoing reasons. would

For the proof” its on carry burden of attempt not even did 10 ppm to benzene at exposure whether question threshold Ante, 653. at health. risk to human significant presents car- successfully is whether OSHA question my view, ques- That of record evidence. on the basis its burden ried First, is there issues. principal to two in turn reduces tion determination supporting OSHA’s evidence substantial per- imprecise are too techniques quantification available If not, then of risks? numerical estimate amit reasonable on the rests regulation show its has failed to OSHA finding Second, is OSHA’s -evidence.” “best available by sub- supported levels current risks at significant show that has failed not, If then OSHA evidence? stantial safe provide necessary to reasonably regulation is new workplaces. and healthful

II disagree I do not close, question regard I Although this rec- failed, has OSHA view that plurality’s sum- issues on the threshold proof carry its burden of ord, to properly OSHA if one assumes that even But marized above. J., 713-714 697-701, (Marshall, post, see met burden, requires also the statute that I conclude dissenting), its standard effects economic that to determine agency An benefits. expected relationship to reasonable bear necessary” “reasonably neither health standard occupational expendi- for statute, if it calls required “feasible,” as nor health expected to the disproportionate wholly tures safety benefits. but ac- only permits (5) not (b)6§ that contends

OSHA health reduce standards promulgate toit requires tually effects those unless effects, to economic regard without risks 668

would cause widespread dislocation throughout an entire in- dustry.4 Under the threshold test adopted by plurality today, this authority will exist only with respect “sig- nificant” risks. But the plurality does not reject OSHA’s claim that it must reduce such risks without considering eco- nomic consequences less serious than massive dislocation. my view, that claim is untenable.

Although one might, wish that Congress had spoken with greater clarity, the legislative history and purposes of the statute do not support OSHA’s interpretation of the Act.5 4 argues OSHA (b) (5) 6 requires it to promulgate standards that are “feasible” only in the sense they “capable are of achievement”; is, achievable “at bearable cost with available technology.” Brief for Federal Parties 57. The lower courts have indicated that a standard is not “infeasible” under OSHA’s test unless it would precipitate “massive economic dislocation” in the industry. affected See, g., e. American Fed eration Labor v. Brennan, F. 2d (CA3 1975). In this case, simply OSHA asked a consulting firm to ascertain the costs of complying with a 1 ppm standard. ante, See at 621. OSHA then con cluded that “the economic impact of [compliance] will not . . . threaten the financial welfare of the affected firms or general economy.” the. Reg. Fed. (1978). The cost of complying with a may “bearable” and still not reasonably related to the expected. benefits A manufacturing company, for example, may have financial resources that enable it pay the OSHA-ordered costs. But expenditures unpro ductive purposes may seriously limit its ability financial competi remain tive and provide jobs. 5 I will repeat the detailed summary of legislative history con tained in the plurality opinion. Ante, at Many 646-652. of the consid erations that plurality relies upon to Congress’ show concern sig nificant persuade harms me Congress did not intend OSHA to reduce significant each hazard regard without consequences. economic Senator Williams, a sponsor of the legislation, stated: “Our bill is fair and reason able. It good-faith is a effort to balance the need of workers to have a healthy sa[f]e work against environment the requirement of industry to function without undue interference.” 116 Cong. Rec. (1970), Legislative History Occupational Safety and Health Act of 1970 (Committee compiled Print for the Senate Committee on Labor and Public Welfare), p. There could be no such “balance” if OSHA were

669 Congress intended believe simply to that unreasonable It to workplaces of risk-free goal desirable pursue the OSHA to indus- viability particular economic that the extent As the thereof —is threatened. segments significant tries —or carry out chosen to itself not OSHA has observes, plurality Ante, at 650. If policy all instances. self-defeating such a of American ability impair the regulations would did, it OSHA and foreign businesses effectively with compete to industries American workers.6 provide employment to in- Congress lightly assume I would not therefore to risks found health require reduction of tended OSHA indus- affected whenever finds also significant consequences regard to economic impose without standards authorized serious dislocation. short of (5) (b) 6 preliminary version a § Dominick described Senator follows: we are say that when . trying in the bill . . was were to do we “What steps as ought to take such agents, agents physical we dealing with toxic per- a which atmosphere within an practical provide and are feasible Unfortunately, we had safety affected. would health or son’s would that no one anyone would be assured language providing that [sic] . . . have a hazard. (1970), Cong. Rec. 37622 . . .” unrealistic standard. “It was an added). (emphasis supra, History, Legislative at fore- objection to the “unrealistic” Dominick’s Senator (8) Act thought imply (b) that he (5) § does not runner of § J., dissent- (Marshall, post, 710-711 at content. See lacked substantive (5) (b) be deleted proposed that hardly have ing) Senator . The other sections thought ante, 647, if he had not entirely, see practical. and reasonable regulations were required health Act task, complex and difficult extremely assigned an OSHA Congress has considerably than clear. less OSHA afforded guidance health title, minimize tois in its responsibility, reflected agency’s primary highly of our health workplace. the economic Yet safety risks adequate employment requires high rate society industrialized be little There can foreign competition. vigorous increasingly response to reasonably societal OSHA balance Congress intended doubt maintaining goal of conflicting often safety in health interest economy. strong national 670

try can bear the costs. See supra. n. 4, Perhaps more sig- nificantly, however, OSHA’s interpretation (b)(5) of §6 would force it regulate in a manner inconsistent with the important health safety purposes of the legislation we construe today. Thousands of toxic present substances risks fairly could be characterized “significant.” ante, as Cf. 645, 51. n. OSHA Even if succeeded selecting grav- in est risks for earliest a regulation, process standard-setting ignored economic'considerations would result in serious misallocation resources and a lower effective safety level of than could be achieved under standards set reference to the comparative benefits available at a lower cost.7 I would not attribute such an irrational intention Congress. cases,

In these OSHA did find that the “substantial costs” of regulations the benzene justified. are supra, See at 665-666. But the record before us contains neither adequate documen- conclusion, tation of this nor evidence that OSHA weighed the relevant considerations. The agency simply an- nounced finding its of cost-justification explaining without by the method which it determines that the benefits justify the costs and their economic effects. No system rational regulation permit can its administrators policy to make judg- ments without how their decisions explaining effectuate the purposes of the governing law. nothing the statute laxity authorizes such in these cases.8 Since neither the air- 7 example, (b)(5) reading For OSHA’s could depletion force the § industry’s of an in an resources effort to single by reduce a risk some amount, speculative though significant even unregulated. other risks remain justify largely policy The decision that costs benefits judgment delegated by Congress. to OSHA When a court judgments reviews such under the "substantial evidence” standard mandated 29 U. S. C. (f), court responsible must determine whether agency has earefulLly] . . ¡ir] reasons win identifi[ed] chooses to follow one another” course rather than most reasonable method of effectuat ing purposes applicable of the law. Dept. Hodg Industrial Union v. 331, son, 339-340, App. C. U. S. D. F. 2d 475-476 identify its Since OSHA failed to reasons cases, these I express no nor the contact stand- dermal concentration borne of the requirements satisfies benzene ard statute, affirming join judgment I the Court's governing Appeals. the Court judgment Rehnquist, judgment. concurring Justice Me. con- present center at the statutory provision Safety Health Occupational of the (b)(5) troversy, §6 Secretary of part, in relevant that 1970, states, Act Labor toxic dealing with standards promulgating . in

“. . set the . shall agents . . physical or harmful materials the extent assures, most adequately which standard evidence, that best of the available on the basis feasible, or of health impairment will suffer material employee no regular has employee if such even capacity functional for standard such with hazard dealt to the S. C. 1594, 29 U. 84 Stat. working life.’’ of his period added). (emphasis (b)(5) petitioners of the is one Secretary, who According to duty, absolute him an upon imposes (b)(5) herein, §6 safe no which for benzene like substances harmful regulating exposure at permissible standard set known, level cost bearable “can achieved level lowest While Parties 57. Federal Brief for technology.” available of “bear- concept refine the attempt Secretary does not proposed that a believes apparently he cost.” able be such “will impact long so its as economically feasible or firms affected welfare financial to threaten as (1978). Reg. 5939 43 Fed. economy.” general agreed, court lower reply, Respondents provision light of another read must be (b)(5) 6§ other appropriate may be review the standard as to opinion situations. Act, (8),

same § which defines “occupational health and safety standard” as

a standard which requires conditions, or adop- tion or use of one or more practices, means, methods, operations, processes, or reasonably necessary appro- priate provide safe or healthful employment places of employment.” Stat. U. C. §652 S. According to respondents, § 6 (b)(5), tempered by 3 (8), requires the Secretary to any particular demonstrate that health justifiable standard is on the basis of a rough balanc- ing of costs and benefits.

In considering these alternative interpretations, my col- leagues a good manifest deal of uncertainty, ultimately divide over whether Secretary produced sufficient evidence the proposed standard for benzene will in any result appreciable benefits at all. This I uncertainty, would suggest, is eminently I justified, since believe litigation pre- sents the Court with what has to be one of the most difficult issues that could confront a decisionmaker: whether the sta- tistical possibility of future deaths should ever be disregarded in light of the economic costs of preventing those I deaths. would also suggest that widely the varying positions advanced parties briefs of the and in opinions the of Mr. Justice The Stevens, Justice, Chief Mr. Justice Mr. Powell, demonstrate, Justice Marshall perhaps better than other fact, that Congress, governmental the body best suited obligated most to make the confronting choice inus litigation, has improperly delegated that choice to the Secre- tary of Labor and, derivatively, to this Court.

I In his Second Treatise of Civil Government, published in 1690, Locke John wrote that power of the legislative, “[t]he being derived people by the a positive voluntary grant institution, can no other than positive what to laws, and not make only to being which conveyed, grant power to can no transfer have legislative legislators, make other hands.”1 it in place making laws of authority their recognized expressly this Court years later, hundred Two ability Congress’ necessity for limits of and the existence Executive of the authority representatives delegate its power legislative delegate cannot Congress “That Branch: as vital recognized universally principle ais President government of system of maintenance integrity and 649, Clark, 143 U. S. Field v. Constitution.” by the ordained (1892).2 how- not, power is legislative of delegation against rule The exception. no for to allow as principle a cardinal ever, so saw statesmen, who practical were Constitution Framers a two-sided powers of separation of doctrine exam- 48, for Paper No. Madison, Federalist James coin. among authority the division while recognized ple, principle, was a useful government branches various essen- requires, maxim which separation degree “the duly practice never can government, free to a tial Lodge ed. (H. p. 308 48, No. The Federalist maintained.” 1888). sealing-off hermetic recognized has also Court

This could another from one government branches the three Government National of a establishment frustrate easily Tradition Government, of Civil Treatise Locke, Second 1 J. treatise, Locke 1957). In the same (M. Mayer ed. Freedom, p. ¶ making laws power of transfer cannot legislative that “[t]he wrote also people, from the power delegated being but hands; any other Ibid. to others.” it over pass cannot they have it who argument rejected considered had 1812, this Court early as As embargo on a trade to terminate authorizing President the a statute that violating neutral “the ceased two nations those if France Britain discretion much delegated too States” United commerce 382, States, 7 Cranch Brig Aurora v. United See Branch. Executive 386, 388. capable of effectively exercising the substantive powers granted to the various branches the Constitution. Mr. Chief Taft, Justice Hampton the Court in W. writing J. & Co. v. States, United 276 U. 394 (1928), S. noted prac- ticalities of the balance that has tó be struck: rule

“[T]he the actual administration of the government Congress or Legislature should exercise legislative power, the President or State executive, the Governor, the executive power, and the Courts or the judiciary the judicial power, and carrying out constitutional division into three branches it is a breách of the National fundamental law if Congress gives up its legislative power and transfers toit or President, n to the Judicial branch, or byif law it attempts to invest

itself or its members with either executive power judi- power. cial say This not to that the three branches are not co-ordinate parts of government one each in the field of its may duties not invoke the action of the two other branches in so far as the action invoked shall an assumption of the constitutional field of action of another branch. determining may what do in seeking assistance from branch, another the extent and character that assistance according must be fixed common sense and the gov- inherent necessities Id., ernmental co-ordination.” at 406.

During the third and fourth decades of this century, Court within relatively period short of time struck down several Acts of Congress on grounds they exceeded authority of Congress under the Commerce Clause or under nondelegation principle of separation of powers, and at the same time struck down state statutes because they violated process “substantive” due or interfered with inter- state See commerce. It. generally Jackson, The Struggle for Supremacy Judicial 48-123 When many these decisions later were principle overruled, Congress simply legislative authority

could transfer to the its Exec- my utive fell under cloud. in opinion a Yet decisions such Refining Ryan, (1935), Panama Co. v. S. 388 suffer U. judicial of of policymaking from none the excesses many of the other decisions of era. The plagued some upheld delegations of congressional later decisions that have on authority largely to the Executive Branch have done so in may authority its theory Congress wish to exercise technical, sufficiently field, but because the field is particular Members sufficiently large, covered and the ground to be in area necessarily expert Congress themselves not may be asked they legislate, choose the most that which lay Congress doctrine is that separation-of-powers under the law, that animate the general policy and standards down the “fill in the standards, those leaving agency to refine cases. These particular blanks,” or standards to apply the above-quoted my mind, simply illustrate the decisions, Mr. ago by Chief Justice years more than principle stated judged must be legislative authority delegations of Taft necessities the inherent “according to common sense and co-ordination.” governmental prin- of these light here legislation at issue Viewing the literally, Read pass fails to muster. I ciples, believe that completely precatory, (b) (5) §of portion the relevant stand- protective Secretary adopt the most admonishing the he cannot. duty if excusing him from that can, if he but ard level a “safe” for which substance of hazardous In the case (5) (b) §of 6 language impractical, or unknown is either con- where indication absolutely no gives the Especially his line. draw safety he should relative tinuum I have stake, at the interests importance light alone, would standing provision issue, doubt no legisla- delegations against uncanalized violate the doctrine then, is whether remaining question, For me the power. tive his- legislative ascertainable are standards additional whether not, if (b) (5) or, statutory context tory such a standardless delegation was justifiable in light of the “inherent necessities” of the situation.

II One of the primary sources looked to in adding Court gloss to an otherwise grant broad legislative authority is legislative history of the in question. statute opin- The ions of Mr. Justice Stevens and Mr. Justice how- Marshall, ever, give little more than tipa of the hat to the legislative origins of 6 (b) (5). Such treatment is perhaps understand- able, since the legislative history of that section, far from shedding light on important what policy Congress choices was making the statute, gives one feeling of viewing the congressional purpose “by the early light.” dawn’s precursor of § 6 (b)(5) was placed in the Occupational Safety and Health Act of while that bill was pending in the House Committee on Education Labor. At time, the section read:

“The Secretary, promulgating standards under this subsection, shall set the standard which most adequately on assures, the basis of the best professional available evi- dence, that no employee will suffer any impairment of health, or functional capacity, or diminished life expect- ancy even if such employee regular has exposure to the hazard dealt with by such standard for the period his working life.” §7 (a)(4), H. R. 16785, 91st Cong., 2d Sess., 49 (1970), Legislative History of the Occupational Safety and Health Act of 1970 (Committee Print com- piled for the Senate Committee on Labor and Public Welfare), p. (1971) (hereinafter Leg. Hist.). aspects Three of this original proposal are particularly sig nificant. First, and perhaps most importantly, as originally introduced the provision contained no feasibility limitation, providing instead that Secretary “shall set the standard which most adequately assures” that no employee will suffer pro Second, have required would harm. or functional of health “any” impairment employees tect face, in its Third, although perhaps its on capacity. safety stand both health and applied to provision intent, Act.3 under the promulgated ards that, journey its point at this doubt There can be little (b)(5) required have the Sec- §6 through Congress, permissible substances, to set the toxic retary, in regulating known, if no safe level or, a safe level level of Public Labor and Committee the Senate at zero. When re- all identical almost provision Welfare considered objected however, Javits Senator version, House spects to the require to “might interpreted question provision that the cases, of feasi- regardless all safety in health absolute Hist. Leg. (1970), p. 58 91-1282, No. Rep. S. bility. . .” . 418. (1970), Leg. Hist. 37327 Cong. 116 Rec. See also 197. that provide bill to therefore amended The Committee adequately which most the standard “shall set Secretary any im- suffer employee would no assured feasibly” Sess., 39 p. Cong., 2d 91st of health. S. pairment added). only addi- (emphasis 242 Hist. (1970), Leg. Senate appeared in change for this explanation tional There, the floor. bill to Senate accompanying Report explained: Committee (b)6 shall under section promulgated

“[Standards appropri- which, where requirements, represent feasible demonstra- research, experiments, on ate, be based shall scientific latest available and the experience, tions, past 3 application, seemingly general that, despite argue its Respondents hazards only health (5) actually (b) referred 6 original version § Respondents B Brief for safety See Addendum opposed to hazards. as proposi support of this Institute et al. 5b-6b. American Petroleum Com history the House legislative where they portion of the tion, cite a 6 proposed version of § and Labor stated Education mittee “occupational health stand Secretary set apply when (b) (5) would Leg. (1970), 848. Hist. 91-1291, p. 18 Rep. H. No. R. ard.”

data. should be assuring, Such standards directed at so as no possible, employee impaired will suffer far capacity expect- health or functional or diminished life ancy, by involved, reason of to the hazard even though such exposure may period be over his en- working Rep. tire life.” S. 91-1282, p. (1970), No. Leg. (emphasis added). Hist. 147 Despite Senator Javits’ inclusion of the words “and feasi- bly” in the provision, participants in floor debate immedi- ately (b) (5) characterized “to requiring utopia any establish a free from hazards” and to “assure there will not be risk Cong. at alb” Rec. (1970), Leg. (remarks Hist. 480-481 Dominick). Sen. Senator Saxbe stated: saying

“When we come to employer guar- that an must employee antee such an protected any possi- I ble it will harm, think be one of the difficult most areas going we are have ascertain. . . .

“I believe the terms that we are passing back and forth are going have to be Cong. Rec., identified.” 26522, Leg. Hist. 345.

In response concerns, to these Senator Dominick introduced proposed substitute for deleting at provision, the sentence entirely. issue here explained He that his amendment delete (b) requirement

“the in section 6 (5) the Secre- occupational tary safety will establish and health stand- adequately feasibly most and which ards assure to no employee will possible impair- suffer extent capacity, or functional ment of health diminished life employee regular if the has even expectancy with the standard for period dealt hazard to the working life. his inherently confusing and requirement “This unrealis- require the Secretary be read ban all tic. It could in which there remains some risk of injury, occupations In health, expectancy. or life the case all impaired impossible will all risks to it eliminate occupations, Thus, could, criteria if present safety health. every business in nation. close this literally applied, cases, might which most many addition, in elimination ‘feasibly’ assure the ‘adequately’ occupation of the prohibition be the danger would itself. more than as no is intended provision

“If the duty, it seems Secretary to his do admonition to included advisable if deemed unnecessary could, original.) (Emphasis history.” legislative 36530, Leg. Hist. 367. Rec., Cong. settled supporters and his Dominick

Eventually, Senator re- agreement (b)(5). This language §of present for the provi- original version from the changes in three sulted was provision First, the by Senator Javits. sion amended standards only to applied explicitly that to state altered apparent agents,” physical or harmful “toxic for materials no Secretary was Second, the safety standards. contrast impair- “any” employees protect admonished to longer impair- “material” only from but health, rather ment their purposes, our importantly Third, and most ments. revamped feasibly assures” adequately “most phrase feasible.” assures, the extent adequately “most to read inter- different number of with a presented have been We Secretary, Senator According to the shift. pretations seemingly delete could not he recognized Dominick instead entirely, and (b) (5) §of requirements absolute or harmful to toxic materials application limit its agreed only to Secretary was specify and to agents physical health. of their impairment material from employees protect set mandate his Secretary asserts Significantly, eco- technologically level safest standards such nomically achievable remained unchanged by the Dominick amendment. According the Secretary, the change in lan- guage from “most adequately and feasibly assures” to “most adequately assures, to the extent feasible,” represented only a slight shift emphasis, perhaps suggesting preference “a for health protection over cost.” App. to Brief for Federal 7a, Parties n. 2. See also Brief for Federal Parties 59. Mr. reads this history quite differently. Justice Marshall In his view, the version of (b)6§ (5) that reached the Senate floor did not “clearly embod[y] the feasibility requirement” and thus was soundly criticized as being unrealistic. See post, at 693. It was only as a result of the floor amendments, replaced which “most adequately and feasibly assures” “most adequately to the assures, extent feasible,” that Secretary clearly was authorized to reject a if proved technologically or economically infeasible. See also post, at 710, and 720-721, n. 34.

Respondents yet cast light a third on these events, focusing upon a places few legislative history where the words “feasible” and “reasonable” were used more or less inter- changeably. See S. Rep. No. 91-2193, pp. 8-10 (1969), Leg. Hist. 38-40; 115 Cong. Rec. 22517 (1969) (Sen. It Javits). is their contention that, when Congress said “feasible,” it meant cost-justified. According to respondents, agree who regard this with the Secretary, the of the meaning feasibility requirement did not change substantially between the version left the Senate Committee on Labor and Public Welfare and the version that was ultimately adopted part as Act. my

To mind, there are several lessons gleaned to be cryptic somewhat legislative history. First, pointed out to the extent Mr. that Senator Justice Marshall, Javits, Dominick, Senator and other Members were worried about imposing upon the impossible burden of assuring safety, absolute they did not view 3 (8) of the Act

681 as a limitation on I duty. that therefore find it difficult to accept court, the conclusion of the lower as by embellished respondents, (8)3§ as a general upon acts check Secretary’s (b) duty (5) under 6 adopt protective § to the most standard feasible. I

Second, and importantly, legislative more believe that the history feasibility requirement, demonstrates as 6 employed in (b)(5), legislative miragé, appearing § is a to some but not to others, assuming any Members and form by desired accept the beholder. am unable to I Justice Mr. argument by that, changing phrasing Marshall’s (b)(5) adequately feasibly §6 from “most to and assures” adequately feasible,” “most to the extent the Senate assures, injected already something into section that was not I regard, there.4 If am correct in then -the amendment Secretary introduced Javits to relieve the Senator duty workplace Dominick create a left Senator risk-free grounds. object provision free to to the amended on the same descrip- Perhaps aptest Senator himself offered Dominick ad- feasibility requirement tion “no more than an as Cong. Secretary duty. monition to do . his . .” (1970); Leg. Rec. 36530 Hist. 367. sum, legislative history nothing indicate contains other language anything “to feasible” does the extent history legislative strongly him that Senator Dominick indicates little, adequately any, phrases “most

self saw if difference between the feasibly assures, adequately extent feasible.” assures” and “most (b) attempt of his first sentence In the course earlier to delete the (5) of that as entirely, paraphrased the unamended version section he “most, adequately requiring promulgate standards possible” feasibly suf assure to the extent employee would that no (emphasis (1970), Cong. Leg. Hist. 367 fer harm. See Rec. 36530 added). between significant difference Unless Senator Dominick found a “possible” difference “feasible,” it clear that there is little the words section perception what the unamended between Senator Dominick’s required after his required way feasibility what that section amendment.

than render what had clear, been a if somewhat unrealistic, largely, if entirely, precatory. There certainly nothing to indicate that these words, used in (b) § (5), are limited to technological and economic feasibility. When Con- gress has wanted to the concept limit feasibility fashion, it has so, said as is evidenced in a statute enacted the same week provision as the at issue I here.5 question also whether wants to assume the duties such an interpretation would impose upon him. In cases, these for example, the Secretary actually declined to adopt lower than 1 ppm for some industries, not because it was economically or technologically infeasible, but rather because “different levels for different industriés would result serious administrative difficulties.” 43 Fed. Reg. (1978). See ante, also at 650 (plurality opinion). If 6 (b)(5) § author- izes the Secretary reject protective a more standard in the interest of administrative I feasibility, have little doubt that he reject could such standards reason whatso- ever, including even political feasibility.

Ill prior In eases this Court has looked to other than sources legislative history to breathe life into vague otherwise delegations legislative power. Light American Power & SEC, Co. v. 329 U. S. 90, (1946), for example, this Court concluded that certain seemingly vague delegations “derive [d] much meaningful purpose content its Act, background factual statutory which they context in appear.” Here, however, there is little nothing or (c)(2)(A) (B) 5 Sections 211 Act, of the Clean Air as amended on 31, 1970, Dec. 84 Stat. authorize the Environmental Protection Agency control, regulate, prohibit or automotive additives after fuel technologically economically “consideration of other means of feasible achieving (c) (2) (A) (1976 emission . . .” standards. U. S. C. § ed., II) Supp. (emphasis added). remaining provisions of the Occupational Safety and Health Act to provide specificity to the (b) feasibility § criterion in 6 true, may It as suggested by Marshall, Mr. Justice the Act as a whole expresses a distinct preference for safety over dollars. But expression of preference, Ias read it, falls far short of the proposition that the Secretary must eliminate marginal or insignificant risks of material harm right down to an industry’s breaking point.

Nor are these eases like Lichter v. United States, 334 U. S. 742, 783 (1948), where this upheld Court delegation of author- ity to recapture “excessive profits” in light of a pre-existing practice. administrative Here, the Secretary’s approach to toxic substances like benzene could predated have (b) enactment of (5) itself. there are Moreover, indica- tions postenactment that the practice administrative has been less than uniform. For example, Safety the Occupational Health Review Commission (OSHRC), body.charged adjudicating citations issued the Secretary under Act, apparently agree does not with the definition of “feasibility,” advanced in these cases by Secretary. In Continental Co., Can 1541, OSHC 1976-1977 OSHD 21,009 (1976), ¶ the Commission reasoned: employers

“Clearly, have finite resources for use available to abate health just hazards. And clearly they if are to be spend made to without limit for abatement of this their hazard financial ability to abate other in- hazards, cluding life threatening Id., hazards, reduced.” at 1547, 1976-1977 OSHD, p. 25,256.

Furthermore, record these cases contains least one Secretary indication that time, quite himself at one was, 6 (b) (5) uncertain what limits placed upon him. an- nouncing the proposed ppm standard discussing its eco- nomic ramifications, explained that “[w]hile precise meaning of feasibility not clear from Act, it is OSHA’s view may term include the economic ramifi- cations of requirements imposed by standards.” 43 Fed. Reg. This candid and tentative statement falls far short of the Secretary’s present position that economic and technological considerations only set the limits on his duty adopt the most protective standard. Finally, as noted earlier, the Secretary has failed apply present his stringent view uniformly, rejecting these cases a lower standard for some industries on grounds of administrative convenience.

In some cases where broad delegations power have been examined, this Court upheld has delegations those because delegatee’s residual authority particular over subjects of regulation. In United States Curtiss-Wright v. Export Corp., U. S. 307 (1936), upheld this Court a statute author- izing the to prohibit President the sale of arms certain countries if he prohibition found that such a would “contrib- ute to the peace.” reestablishment of This Court reasoned that, the area foreign affairs, Congress “must often ac- cord to the President a degree discretion and freedom statutory restriction which would not be admissible were domestic affairs alone Id., involved.” at 320. Similarly, United Mazurie, States v. 419 U. S. 544 (1975), upheld a delegation broad authority to various Indian tribes to reg- ulate the introduction of liquor into Indian country. Ac- cording to Mazurie limitations on Congress’ authority to delegate legislative power are “less stringent in cases where entity exercising the delegated authority itself possesses independent authority subject over Id., matter.” at 556- cases, 557. In present however, neither the Executive general Branch in nor the particular enjoys any independent authority subject over the matter at issue. earlier, in

Finally, as some indicated cases this has Court by necessity, upholding a rule of abided delegations broad authority it would where “unreasonable impracticable compel Congress prescribe regarding detailed rules” Light American Power & Co. policy or situation. particular S., SEC, v. 105. v. See also Stranahan. U. Buttfield (1904). But no need for such an 470, 496 evasive 192 U. S. “feasibility” apparent present is cases. standard clear, (b)(5), faced with a if drafting Congress § 6 dif- balancing statistical lives and industrial choice between ficult, Secretary to life elevate human authorizing or resources in- dislocation in an affected concerns save massive above all difficulty of choice Congress recognized That this dustry. Saxbe, noted of Senator previously from remark is clear em- saying that an we come who stated “[w]hen employee protected must that such an ployer guarantee I of the most think it will be one possible harm, Cong. to have to ascertain.” going we are difficult areas Congress chose, (1970), Hist. 345. Leg. Rec. |That choice difficult pass this intentionally unintentionally, quality spectral is evident from on to the Secretary un- Saxbe’s in Senator capsulized it selected land back passing we are “the terms that promise fulfilléd Ibid. to be identified.” going forth are to have IV nondelegation Court, the by As and enforced formulated most First, functions. important serves three doctrine orderly consistent extent abstractly, it ensures to social choices important administration governmental our Government branch of Congress, the made are policy California, Arizona v. will. See popular responsive to most in part); dissenting (Harlan, (1963) 546, J., S. 373 U. (1967) (Brennan, Robel, 389 U. S. v. United States guarantees result). Second, doctrine concurring in J., necessary delegate finds it Congress extent that, to the with an authority recipient provides authority, “intelligible principle” guide delegated exercise of the discretion. See J. Hampton States, W. & Co. v. United S., S., 409; U. Refining at Panama Ryan, Co. v. 430. U. Third, second, and derivative of the the doctrine ensures that courts charged with reviewing the exercise of delegated legis lative discretion will be able to test against that exercise as certainable standards. See California, Arizona v. supra, at 626 (Harlan, J., dissenting part); in Light American Power & SEC, Co. supra, v. at 106.

I believe legislation at issue here fails on all three counts. The decision whether law of diminishing returns should have place in regulation of toxic quin- substances is tessentially legislative one of policy. For Congress pass that decision on to violates, in the manner it did my mind, John Locke’s caveat —reflected the cases cited opinion legislatures earlier laws, are to make —that legislators. I Nor, prior as think amply discussion demon- strates, do provisions history at issue or their legislative provide the Secretary any guidance him might lead to his somewhat tentative conclusion he must eliminate economically as far technologically benzene as possible. I Finally, would suggest the standard of “feasibility” meaningful judicial renders impossible. review ought shy

We away judicial duty not to from our to invali- delegations solely date legislative authority unconstitutional thereby out concern that should reinvigorate we discredited constitutional pre-New doctrines of the Deal If non- era. delegation doctrine into has fallen same desuetude as have process interpretations substantive due restrictive it, is, phrased Commerce one writer has Clause, it “a case Ely, Distrust, of death association.” J. Democracy and Indeed, A Theory Review 133 Judicial a number suggested observers have this Court should once more up ensuring Congress take its burden not un- does delegate important necessarily po- of social policy choices

687 observers, liticálly unresponsive administrators.6 Other as might imagined, disagreed.7 have

If we are ever reshoulder the burden of ensuring decisions, Congress itself make the policy critical these are surely the cases which to do it. It is difficult to imagine a example more obvious Congress simply avoiding a choice which was both purposes fundamental for of the statute and yet politically so divisive that the necessary decision or com- difficult, promise if not impossible, to hammer out in the legislative forge. Far from detracting from the substantive authority Congress, a declaration that the first sentence of (b) (5) 6 Occupational Safety § of the and Health Act consti- delegation tutes an invalid of Labor would preserve authority Congress. If Congress wishes legislate in an area which it has not sought previously to en- ter, it will in today’s political undoubtedly world run into no opposition legislation matter how the is formulated. But very legislative authority that is the essence of sys- under our tem. It is the hard choices, and not filling of the blanks, which must be made the elected representatives people. policy When fundamental decisions under- lying important legislation about enacted are to be made, the buck stops with Congress and the President insofar his as he exercises constitutional role in legislative process. I would (b) invalidate the (5) first sentence of 6 Safety Occupational and Health Act of 1970 it applies Democracy Distrust, Ely, See Theory J. A of Judicial Review (1980); Freedman, 131-134 J. Legitimacy, Crisis The Administra (1978); tive Process and American Lowi, Government 78-94 T. End Ideology, Policy, Authority of Liberalism: and the Crisis of Public 129- 146, (1969); Beyond Wright, Discretionary Justice, 297-299 81 Yale L. J. 575, (1972); Waist-Deep Post, 3, Regulation, Washington 582-587 Nov. 1979, p. A10, Douglas, East, Young (1974). col. Cf. 1. W. Go Man 217 Davis, Discretionary Preliminary Inquiry K. A See Justice: 49-51 (1969); Stewart, Law, The Reformation of American Administrative 1669, (1975). Jaffe, L. Rev. 1693-1697 Cf. The Illusion of the Harv. Administration, 1183, L. Ideal Harv. Rev. n. any toxic substance or harmful physical agent for which safe level, is, a level at which “no employee will suffer mate- impairment rial of health or functional capacity even if such employee has regular exposure to period [that hazard] life,” of his working is, according to the Secretary, unknown or otherwise “infeasible.” Absent further congressional action, the Secretary would then have to choose, acting when pursuant (b) to 6 (5), § between a setting safe standard or set- ting no standard at all.8 for the stated Accordingly, reasons I above, concur in judgment the of the affirming Court the judgment of the Court of Appeals.

Mr. Justice Marshall, with whom Mr. Justice Brennan. White, and Mr. Mr. Justice join, Justice Blackmun dissenting.

In cases of statutory authority this Court’s construction, If limited. the statutory legislative language and intent are judicial plain, the inquiry is at an end. juris- Under our prudence, is presumed that ill-considered legisla- or unwise tion will be through corrected the democratic a court process; permitted is not to distort a statute’s meaning order to make it conform with the Justices’ own views of sound social policy. TVA Hill, See v. 437 U. S.

Today’s decision flagrantly disregards these restrictions on judicial authority. The plurality ignores plain the meaning Occupational Safety and Health Act of order bring authority of the Secretary of Labor in line with plurality’s own views of proper regulatory policy. The consequence unfortunate is that the Federal Government’s to protect efforts American workers from cancer and other crippling may substantially impaired. diseases be ruling upon This would have effect governing standards toxic feasible, physical agents substances or which safe harmful levels are upon extant promulgated standards “national consensus standards” Secretary’s (a), upon authority promulgate under nor “emer §6 (c). gency temporary under standards” provides: Act of (b)(5) § 6 of sentence first with dealing standards promulgating Secretary, in “The under agents physical harmful or materials toxic adequately most which standard set subsection, shall best of basis on the feasible, extent to the assures, material suffer will no evidence, employee available if such even capacity functional health or of impairment dealt hazard exposure regular has employee life.” working his of period for the such (b)(5). §655 C. U. S. of basis on the found, Labor case

In this creates benzene exposure (1) evidence, substantial non- variety damage, and chromosomal cancer, risk even disorders, blood fatal potentially but malignant shown; been has level safe no (2) ppm; 1of level derived be would lives saved form benefits (3) lives that number (4) standard; permanent relatively substantial either out turn could saved it knowledge, of scientific state present the (5) under small; number the way rough in a even calculate impossible is assumptions making without least saved, at would that lives commu- medical much absurd appear would harm materially not would the standard (6) nity; and does Court The industries. covered condition financial not could it Thus, findings. these aside set not in accord fully Secretary’s decision plainer . . . assur[e] [to] adequately “most mandate statutory his health impairment material suffer will employee no . .” . capacity. functional or its based is contrary conclusion plurality’s occu- defines which (8), 652§ S. C. TJ. 29of interpretation requires “which one standard health safety pational provide appropriate or necessary reasonably . . . conditions plu- According .” . . employment. healthful safe appropriate” necessary or “reasonably is rality, unless the Secretary able to show that “at least more likely than not,” ante, at 653, that the risk he seeks to reg- ulate is a “significant” one. Ibid. Nothing the statute’s language or legislative however, history, indicates that “reasonably necessary or appropriate” language should be given this meaning. Indeed, both demonstrate that plu- rality’s standard bears no connection with the acts or inten- tions Congress and is based only on the plurality’s solicitude for the welfare of regulated industries. And plurality uses this standard to’ evaluate not the agency’s decision in case, but a strawman of its own creation.

Unlike the I plurality, do not purport to know whether the actions taken by Congress and its delegates to ensure occu- pational safety represent sound or unsound regulatory policy. The critical problem in cases like ones at bar is scientific uncertainty. While has science determined to benzene at levels above ppm creates a definite risk of health impairment, magnitude of the risk quan- cannot be tified at the present time. The risk at issue hardly has been shown to be insignificant; indeed, future research may reveal the risk is in fact considerable. But the evi- existing dence may frequently be inadequate to enable the to make the threshold finding of “significance” that the Court requires today. If so, the consequence of the plurality’s ap- proach would be to subject American workers ato continuing risk of cancer and other fatal diseases, and to render the Fed- eral powerless Government to take protective action on their behalf. Such an approach place the burden of med- ical uncertainty squarely on the shoulders of the American worker, the intended beneficiary of the Occupational Safety and Health Act. It is fortunate indeed that at least a major- ity reject Justices the view the Secretary is pre- vented from taking regulatory action when the magnitude of a health risk cannot quantified on the basis of current ante, techniques. See at 666 (Powell, J., concurring in part

691 656, ante, and 63 at n. see also judgment); concurring and opinion). (plurality Act, and be- in the basis holding has no today’s

Because regulatory its own authority impose to has no Court cause the Nation, I dissent. policies on

I Health and Safety Occupational enacted Congress grim “the as was characterized to what response as a Act needs health occupational failure heed history our legis and voluntary action failure The workers.” of our 91-1282, (1970), 4p. No. Rep. see S. level, the state at lation “worsening” and in a “bleak” 144, resulted had Hist. Leg. a result annually as had persons died 14,500 in which situation preceding years four workplace. in the of conditions work in the killed were Americans more passage, Act’s No. Rep. War, S. Vietnam contemporaneous place than as “a designed The Act 91-1283, 142. Leg. 2, Hist. 3 Its workers.” million to 60 for close rights safety bill of every working possible so far “to assure purpose stated working healthful and safe the Nation woman man S. C. 29 U. resources.” human our preserve conditions Safety Occupational Roofing v.Co. Atlas See (b). 651§ 444-445 Comm’n, S.U. Review Health provisions. two through primarily is enforced Act furnish employers upon imposed duty” is “general First, a recognized “free employment places employment or serious death likely cause or are causing that are hazards Second, the (a)(1). §C. 654 29 U. . . .” S. harm. physical safety “occupational set Secretary of Labor authorized Act of Safety and Health Occupational History Legislative Public Labor and Committee for the compiled Senate Print (Committee Leg. (hereinafter Williams) by Sen. (Foreword (1971) Welfare), p. iii Hist.). Leg. 142. (1970), Hist. p. 2 91-1282, No. Rep. 2 S. iii. Leg. Hist.

692

and health standards,” defined as standards requiring “con- ditions, or the adoption or use of one or practices, more means, methods, operations, or processes, reasonably neces- sary appropriate to provide safe or healthful employment places of employment.” 29 U. C. S. 652 legislative history the Act Congress’ particu- reveals lar concern for health hazards of “unprecedented complexity” that had resulted from chemicals whose toxic only effects “are now being discovered.” S. Rep. No. 91-1282, supra, Leg. Hist. 142. “Recent scientific knowledge points hitherto unsuspected cause-and-effect relationships between occupational exposures many of the so-called chronic cancer, respiratory ailments, allergies, disease, diseases— heart and others.” Ibid. of Congress Members made repeated references to the dangers posed by carcinogens and to de- fects our knowledge of their operation and effect.4 One of the primary purposes of the Act towas regulation ensure 5 these “insidious ‘silent’ killers.” special

This concern led to the enactment of the first sen- tence of 29 U. C. §655 S. (b)(5), which, as noted above, provides:

“The Secretary, in promulgating dealing standards 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 regular has exposure to the hazard dealt with by such standard for period of his working life.” This directive is designed to implement legislative three pur- Rep. 4 S. 91-1282, p. No. 2 (1970), Leg. 142; Hist. Cong. Rec. (1970), Leg. (Sen. Hist. 415 Williams); H. Rep. R. 91-1291, No. p. (1970), Leg. 849; Hist. 116 Cong. Rec. (1970), 38392-38393 Leg. Earth). Hist. (Rep. Cong.

5 116 (1970), Rec. Leg. (Sen. Hist. 1003 Daniels). sub- may be there that recognized First, Congress poses. or fre- repeated upon only dangerous become that stances required to was therefore Secretary The exposure.6 quent mate- cause would that substances even protection provide an throughout occurring only upon impairment rial the Second, requirement the working life. employee’s was evidence” available “the best of basis on the Secretary act not would process standard-setting the ensure to intended Recog- views. of scientific uncertainty by the destroyed Congress inadequate, may be knowledge existing nizing that informa- definitive until wait Secretary to the require not did the intended not “it Thus obtained. could be tion medical surrounding diverse by debate paralyzed Secretary be Leg. Hist. (1970), 18p. 91-1291, Rep. No. H. R. opinions.” killers” “silent for concern special Congress’ Third, 848. Secre- strong directive especially justify felt was 37622 Rec. Cong. standard-setting process. tary Dominick). (Sen. Leg. Hist. (1970), not however, is (b)(5), by §655 authority conferred The limita- primary two contains itself subsection The absolute. was impairment “material” requirement tions. substances regulating Secretary from prohibit designed Moreover, employees.7 affected hazard a trivial create must subsection under promulgated standards all con- expressed Congress debates floor During the “feasible.” embodying clearly bill, version prior cern require feasibility requirement, risks eliminate order to industries whole down close unrealistic.8 criticized This impairment. No. Dominick); R.H. (Sen. 37623, Hist. Leg. Rec., at Cong. 6 116 858. Leg. (1970), Hist. 91-1291, 28p. n. See infra. provided: had bill version An earlier subsection, shall under standards promulgating Secretary, “The basis assures, on feasibly adequately most which the standard set impairment suffer will employee no evidence, that available the best The feasibility requirement was imposed as an affirmative limit power. standard-setting The remainder §of 655 (b)(5), applicable to all safety health standards, requires the Secretary to base his standards “upon research, demonstrations, experiments, and such other information as may be appropriate.” In setting standards, the Secretary is directed to consider “the attainment of the highest degree of health and safety protection employee” and also “the latest available scientific data in the field, the feasibility the standards, and experience gained under this and other health and safety laws.”

The Act makes provision judicial for review of occupational safety and health standards promulgated pursuant to § 655 (b)(5). The reviewing court uphold must the Secretary’s of health or functional capacity, or diminished life expectancy even if employee such regular has exposure to the hazard dealt such standard period for the working his life.” S. 2193, Cong., 91st Sess., 2d 39 (1970), Leg. Hist. 242. This standard, it feared, “could be require read to the Secretary to ban occupations all in which there remains some risk injury, impaired health, or life expectancy. In the case of all occupations, it impos- will be sible to all eliminate risks safety to Thus, health. present cri- teria could, if literally applied, every close business in this nation. addition, many cases, the might which 'adequately’ most 'feasibly’ assure the elimination of danger prohibition the occupation itself.” Cong. 116 Rec. (1970), Leg. Hist. (Statement on Amendment of Sen. Dominick). In explaining the present language, Senator Dominick stated: “What we were trying to do in the unfortunately, we did not have bill— proper wording or proper drafting say to that when we are —was dealing with agents toxic physical agents, we ought to steps take such as are practical feasible ánd provide an atmosphere within which per- son’s safety health or would not be affected. Unfortunately, we had lan- guage providing anyone would be assured that no one would have a hazard ... so that no one would have problem for the rest of his working life.

“It was an unrealistic standard. As modified, we would be approaching problem by looking problem at the and setting a standard or criterion which would not result in Cong. harm.” Rec., 37622, Leg. at Hist. 502. evidence by “substantial supported they are if determinations (f). §C. TJ. S. whole.” aas record considered I turn. now evidence It is to that

II case is both in this record discussion plurality’s is ar- It unfair. extraordinarily extraordinarily arrogant own factual make its presumes plurality because rogant relating issues variety disputed ato respect findings with 656-657, n. ante, g., See, e. at regulation. carcinogen Members of necessary to remind not be It should 64. independ- to undertake appointed not they were Court made findings scientific supported adequately review of ent discus- plurality’s And the expert agency.9 technically by a Secretary’s its characterization because is unfair sion what the no resemblance practically bears report sugges- plurality’s to the Contrary case. in this actually did Draconian blindly on some rely Secretary did tion, If he 635-636. 624-625, ante, See carcinogen “policy.” observed him have sufficient been would have it had, court federal for a appropriate course, suggest not, of 9 I do findings fact agency’s blindly to defer reviewing agency action Park, Overton to Preserve Citizens policy. Under determinations and a “search undertake (1971), courts must 402, 416 U. S. Volpe, 401 Inc. v. inquiry is Such factors. inquiry into those judicial ing careful” and Kleppe v. Sierra look,” “hard a to take agency require the designed to considering by omitted), (1976) (citation 390, 410, n. Club, S. 427 U. There manner. reasonable a them in weighing proper factors under decisions scrutiny agency judicial rigorous especially for room also Co., Products v. Carolene States United offered akin a rationale v. Fund Environmental (1938). See 152, Defense 144, n. S. 304 U. 74, 2d 584 439 F. App. C. S. D. Ruckelshaus, U. today, plurality by the taken however, approach basis, no I see reg- fact and of questions of review nearly novo de amounts which by unable no means that are institutions on behalf ulatory policy inap- especially review Such process. political themselves protect which about ones are issue questions at factual when propriate expertise. have expected to reasonably be cannot Court benzene is a carcinogen, proposition respondents do dispute. Instead, the Secretary gathered over 50 vol- umes of exhibits and testimony and offered a detailed evenhanded discussion of the relationship between exposure to benzene at all recorded exposure levels and chromosomal damage, aplastic anemia, and leukemia. that discussion he evaluated, and took seriously, respondents’ evidence of a exposure safe ante, level. See also at 666 (Powell, con- J., curring in part and in judgment).

The hearings on the proposed extensive, standard were en- compassing 17 days from July 19 through August 1977. The 95 witnesses included epidemiologists, toxicologists, phy- sicians, political economists, industry representatives, members of the affected work force. Witnesses were sub- jected to exhaustive questioning representatives from a variety of interested groups and organizations.

Three basic positions were presented at the hearings. The first position was that the proposed 1 ppm necessary because to benzene would cause material impairment of the health of workers no matter how low the exposure level. Some direct evidence indicated that exposure to benzene had caused chromosomal damage, blood disorders, and leukemia at or below the 10 ppm level itself. More im- portant, was suggested that the recorded effects of benzene higher required levels an inference that leukemia and other disorders would result at levels 1 ppm lower, espe- cially after the prolonged exposure typical in industrial set- tings. Therefore, the standard should be set at the lowest level, feasible which 1was ppm. position second was that a 1 ppm exposure level would pose

itself an unwarranted threat to employee health and safety and that the available evidence signifi- necessitated cantly lower level. An exposure limit below 1 ppm, it was argued, would be feasible. There were suggestions ben- that zene was gradually being replaced in many of the affected already operating at were companies most that and industries 1 ppm level. or below should 1971 standard was position third

The evidence suggested that position of this Proponents retained. uncer- was to leukemia exposure of benzene levels linking low and sufficiently safe, limit was exposure the current tain, insufficient would be standard proposed the benefits was testi- addition, there costs. standard’s justify the standard proposed required expenses mony that prohibitive. would be for permanent announcing regulations

The of rea- statement by an extensive accompanied are benzene hearings. results of evaluating the and summarizing sons exposure showed that evidence Secretary found The of non- variety damage, chromosomal causes benzene Reg. 5921 43 Fed. disorders, leukemia. blood malignant imposed low concentrations He concluded action regulatory sufficiently to call grave that was hazard the Act. under referred The deleterious Evidence effects. could that benzene demonstrated conclusively which

studies Id., 5932. at cells. blood-forming chromosomes damage relationship between a causal testimony suggesting There was not be it could although leukemia, damage chromosomal damage such what extent whether determined suggested had Some studies Id., 5933.10 at health. impair ppm of 10-25 levels exposure damage at chromosomal showing the curve, dose-response No quantitative lower.11 chromo- incidence levels and exposure relationship between Id., 5933-5934. at yet be established. damage, could somal Secretary’s was, damage of chromosomal The evidence Id., 5933. at concern.” for “serious cause view, a a de- of benzene effect most common 258-259, 1039. Tr. 11 Id., 148,200-201, 258. 698

crease in the levels platelets of blood and red and white blood cells. If sufficiently severe, could be pancytopenia result or aplastic anemia, noncancerous but potentially fatal dis- eases. There testimony was some the nonmalignant blood disorders by caused benzene exposure progress could to, or represented, preleukemic stage might eventually which evolve into a Id., frank leukemia. at 5922.12

Considerable evidence showed an association between ben- zene nonmalignant blood exposure disorders at low levels. Such an association had been study established one which the levels frequently ranged from zero to ppm25 some concentrations above 100 ibid.; ppm, they in another ranged from id., ppm, at 5923. Because of the ab- sence of adequate data, a dose-response curve showing the relationship between benzene exposure and blood disorders could not be constructed. There was testimony, considerable however, that such disorders had resulted from exposure to benzene at or near the current level of ppm10 and lower.13 The Secretary concluded that the current standard did provide adequate protection. He observed a “safety factor” to 100 generally to discount used the level at which a causal connection had been found in existing studies.14 Under this approach, he that, quite concluded apart risk, leukemia permissible exposure limit should be set at a level considerably lower than 10 ppm. there was

Finally, substantial evidence that benzene caused leukemia. The concluded evidence established that benzene was carcinogen. A causal relationship between benzene and leukemia was first reported France in since similar time results had been found in a number of countries, including Tur- Italy, Japan, key, Switzerland, the Soviet Union, and the United 12 Id., 145, 173-174, 352, 1227,1928, 3206; Record, at 43B, p. Ex. 166. 13 Id., 149, 360-361, 997, 1023, 2543, 2689, 3203; 11 Record, Ex. 3. 149, 1218, 2692, Tr. 2847. In- National study, undertaken latest States. (NIOSH) Safety Health Occupational stitute incidence the normal over excess fivefold a reported 1970’s, industrial at benzene to exposed among workers of leukemia study seri- testimony that was There in Ohio. plants the risk.15 ously understated low that suggesting studies certain reviewed

The excess any cause not did more ppm10 levels exposure suffered suggested, he studies, Those of leukemia. incidence frankly authors their defects, as methodological severe study a Secretary discussed the Finally, acknowledged.16 at leukemia in excess significant statistically suggesting certain despite that, He found Ibid.17 ppm. 9 2 to levels consistent be considered should study, it in the deficiencies risk leukemia demonstrating an excess other studies Id., at 5928. benzene. exposed employees among areas three Secretary examined uncertainty. The oj Areas Secretary ob 15 the 874, 769-770, 2445. As 768, Id., 747, 314, at extensively study was the NIOSH exposure level served, the the issue Commission Industrial the report from hearings. A during the debated 10 or ranged from zero generally suggested that concentrations of Ohio hearings at evidence that Secretary concluded ppm. But 15 sub sometimes study had period during the exposures area showed that conflicting evidence of the Because level. stantially exceeded risk leukemia excess data, monitoring found he absence of ex any particular be linked study could not in the NIOSH observed posure level. example, heavily, for industry most relied study on which As to the admissions, observed own author’s repeating the Secretary, largely have been may not sample employees included (1) a number followup of inadequate (2) was time; there to benzene exposed leukemia may contracted have who persons employees, so that numerous subject to serious were diagnoses (3) the data; included were unnoticed; (4) deter no gone may have of leukemia question, cases his occupational (5) the made; and had been exposure levels mination Reg. 5928 Fed. admittedly incomplete. were workers tories Record, 154. Ex. ; 22A 1023-1024, Tr. of uncertainty particular that had relevance his decision. First, pointed he latency period evidence range benzene-induced years. leukemia could from to over 20 Id., at exposure 5930. Since lower levels lead to an increase in the latency extremely it would be period, difficult to obtain evidence showing the dose-response relationship leu- between and exposure kemiá to low levels of benzene. Because there has been no adequate monitoring would be past, practically impossible to determine what exposure levels at a sufficiently were time latency distant so period elapsed. problem would have compounded by difficulty of conducting study. suitable Because approaching levels ppm10 had required only been recently, showing direct evidence the relationship between leukemia exposure and levels between and 10 ppm would be unavail- able the foreseeable future.

Second, the Secretary observed that individuals have dif- ferences their susceptibility to Among leukemia. Ibid. exposed those to benzene a group possibly was of unknown but having substantial size various “predisposing factors” whose especially members were Id., vulnerable to the disease. at 5930, 5946. permanent The designed standard mini- was to mize the effects of exposure for susceptible these individuals well as for the relatively insensitive, id., 5946, at also early to facilitate diagnosis Id., and treatment. at 5930. The Secretary discussed the contention that safe level of exposure to benzene been had demonstrated. From the testi mony of numerous scientists, he concluded that it had not. Id., He at 5932.18 also found although that dose-response no id., curve could plotted, be 5946,19 at the extent risk testimony Aksoy, The of Dr. one of the leading experts, world’s was id., typical: ppm . one . . causes “[E]ven leukemia.” Tr. 204. See also at 150, 262, 328, 351-352, 363-364, 394, 1057, 1210, 745-746, ; 2420 9 Record, 2.8-272, p. Ex. 1. 130, 360, 760-761, 414-415, 416-417, Tr. 781-782, 925, 1055-1056; Record, 75, p. 2; Record, 2-4, p. Ex. Ex. 11. Exposure Ibid.20 level. exposure decline would than dangerous less be therefore would 1 ppm a level at Secretary found The ppm. of 10 one at exposure not he should justified the conclusion evidence existing exposed be continued employees while for answers" “wait levels. at hazardous benzene argument responded Secretary Finally, 1 ppm. or lower than zero be level should permissible already had many industries though Even Id., 5947.21 at level would level, lower found that he 1 ppm achieved Ibid. be feasible. dis- a detailed offered The Costs benefits. play should considerations economic the role that cussion must standards He observed his determination. his technologically. “feasible,” economically both under feasible benzene permanent view on the primarily fall would impact economic tests. both petro- refining and petroleum industries, such as stable more industries Id., These 5934. at production. chemical to con- them pass or to the costs readily absorb be able 157,000 industries, involving affected the 20 None sumers. 5935, would id., employees, 629,000 exposed facilities id., He at 5934. expenditures, required bear unable fi- within “well were costs compliance concluded Id., at 5941. industries.” covered capability nancial of the impact economic the national survey of An extensive first-year found contractor, standard, private a undertaken million, recurring $205 $187 between of costs operating con- engineering million, in investment $34 costs annual at- not have respondents Since million.22 $266 about trols 20 2846, 2842-2843. 1372, 405, 382, 401, Tr. should benzene Id., limit exposure ("the permissible at 148-149 seq., et id., at also Aksoy). See Dr. zero”) (testimony of seq. et per employee, expenditure amount of the plurality’s estimate The benzene costs misleading. the Most ante, at 629, highly see tacked Secretary’s the conclusions the Secre- cost, basic tary’s extensive discussion need be summarized here.

Finally, Secretary the discussed the benefits to be derived permanent had During standard. hearings, argued been Secretary should estimate the health proposed benefits of the regulation. To would do this he required to a dose-response construct at least showing, curve rough way, the number of be saved at lives would possible exposure each level. ben- Without some estimate of it was efits, argued, Secretary’s would decisionmaking During be defective. hearings industry witness at- tempted to such dose-response Restricting construct curve. himself carcinogenic effects, he proposed estimated that every years would save two lives suggested six relatively minor justify regu- benefit would not lation’s costs. Secretary rejected

The hypothesis the standard only years. save two lives in six This he estimate, concluded, impossible to was reconcile with the evidence record. Ibid.23 He determined because of numer- that, only standard would be incurred protect onee and would thus an unascer- employees future; tainable number of in the that number will be much higher currently employees employed. than the number of projection, designed extrapolation The amalgamation as an from an existing studies, dependent was on a number assumptions which the Sec retary reasonably questionable. Indeed, could view as the witness himself lousy stated that his estimate on data,” “slightly was based “a set of was basis,” id., guess,” 2772, better than a Tr. and that there was “no real dose-response wholly for a dependent. curve which the estimate was assumptions severely during hearings, The witness’ were tested see id., seq., et Secretary reasonably reject at 2795 and the could them on the example: basis of evidence record. For (1) ap- The witness peared previous to assume that tests leukemia had been contracted exposure; after a lifetime of the evidence afforded no basis for that as- sumption, may quite and the duration of have been short for particular employees. period short, If the duration the witness’ esti- (2) mate been much would have too low. The witness ex- assumed that posure study ppm. were levels in the NIOSH around 100 impossible data, it was existing uncertainties ous those from by extrapolating curve dose-response a construct the Secre- generally, More levels.24 exposure to lower data there was made, and assumption could however, that no such found, 10-15 and zero generally been between had exposure levels that evidence linear was curve dose-response that the assumed (3) witness The ppm. case In the assumption. that for no basis there was levels, but all at promul- Secretary has the carcinogen which (another vinyl chloride of the dose-re- suggested that standards), evidence recent exposure gated steep the levels less becomes and low doses steeply at rises curve sponse the NIOSH in the of workers percent Twenty-five (4) increased. are they still were that assumed witness found, and the been study not had workers hundred additional Six leukemia. not contract would and alive they would too alive; assumed witness the study were still in that exposed that, for these testimony considerable was There leukemia. contract risk. the significantly underestimated study reasons, NIOSH other study found (5) The NIOSH not. had that assumes witness The assumed witness exposure; benzene risk excess fivefold testimony that finding and the NIOSH despite lower, much was excess light of risk. of understatement significant awas finding esti- the witness’ conclude could uncertainties, these unsupportable. was mate dose-response inability construct testifying to Witnesses incidence correlating the of impossibility primarily referred curve levels damage with the disorders, chromosomal leukemia, blood Kraybill Herman Dr. Thus past studies. exposure duration testified: Institute Cancer National many of done, as been This has risk factors. to estimate like “[W]e ago. years several vinyl chloride recall, with you data, experimental basis of] on [the factors the risk estimate . . “. [T]o toxicity say IWhen toxicity data. good you have if presupposes this we indeed chloride, which vinyl data on dose-response good data, I mean that. have did situation, so have didn't we appeared benzene, it “But . . up. . gave therefore, us most of Id., 760-761. at out.” of struck we sort benzene, . “. With . duration levels uncertainties enormous Because arbitrary. necessarily be would assumptions studies, prior conclusion ultimate great so assumptions range possible 1055-1056. 415, id., See entirely uninformative. tary observed that it had not been established that there was a safe exposure level of for benzene. Since there was con- testimony siderable that the risk would decline with the id., exposure level, at 5940, the new standard would save lives. The number of lives saved “may be appreciable,” but there was no way to make a precise more determination.25 question The “on the frontiers of scientific knowledge.” Ibid.

The Secretary that, in light concluded of the scientific uncer- tainty, he was not required to calculate benefits more precisely. Id., at 5941. event he gave “careful consideration” to the question of whether the substantial admittedly costs justified were light hazards of benzene exposure. He concluded that those costs were “necessary” in order to promote the purposes of the Act.

Ill A This is not a case which the Secretary found, or respond- ents established, that no benefits would be derived from a permanent standard, or that likelihood benefits was insignificant. Nor itwas shown quantitative that estimate of benefits could be made on the basis of “the best available evidence.” Instead, the Secretary concluded that benefits will result, those benefits “may” appreciable, be but the dose-response relationship low levels of benzene point At one Secretary did appreciable indicate that benefits were “likely” to result. The Court Appeals held conclusion was unsupported by substantial evidence. Secretary’s The suggestion, how ever, was made in the context of a lengthy discussion intended show appreciable "may” predicted benefits but that their likelihood could quantified. not suggestion The should not be taken as a definitive appreciable statement benefits were probable more than not. infra, For reasons stated there nothing prohibit the Act Secretary acting when he is appreciable unable conclude that are probable benefits more than not. disorders, and blood nonmalignant leukemia, impossible to determine. damage was chromosomal circumstances, whether, in these presented question or whether action, regulatory take permits Act infor- definitive exposure until more continued allow he must becomes available. mation must Secretary’s determinations above, the

As noted record by evidence supported “substantial upheld if stand (f). 655 This 29 C. a whole.” S. U. considered action regulatory judgment legislative a represents ard the traditional stringent more than subject review should be rulemaking. for informal “arbitrary and capricious” arbitrary capricious standard that the have observed We in facts” into the “inquiry searching a contemplates itself aon was based the decision “whether determine order to has there whether factors and of relevant consideration Overton Preserve Citizens judgment.” a clear error been perform Careful (1971). 416 402, 401 U. S. Volpe, Park v. has Congress when important especially of this task is ance evi “substantial rigorous more comparatively imposed judi however, emphasized, have As we requirement. dence” ultimately is test evidence the substantial under review cial S. Perales, 402 U. v. g., Richardson e. See, deferential. Comm’n, U. S. Federal Maritime .; v. (1971) Consolo to the is entitled decision agency’s (1966). The 607, 618-621 not au court is validity, and presumption traditional Secretary. for that judgment its to substitute thorized acted factors the decisional considered Secretary has If the must decision ultimate his statute, with the in conformance 621. Id., respect. large measure given my which, factors to three insensitive plurality health safety occupational review judicial make view, dif- particularly test evidence substantial under the standards technical level high reach often issues First, ficult. required are the courts circumstances In such complexity. unaccus- they are to which matters themselves immerse Second, tomed training experience. the factual issues with which the Secretary frequently must deal are not sub ject definitive resolution. Often “the factual finger points, does not conclude.” Society Industry, Plastics A, v. Inc. OSH 2d 1301, (CA2) F. cert. (Clark, J.), denied, 421 S. U. Causal connections and theo extrapolations retical may be uncertain.' when the Third, question involves acceptable determination of the level of risk, the ultimate decision must necessarily be based on con siderations of policy empirically ‘as well as verifiable facts. Factual determinations can at most define the risk some way; statistical the judgment whether that risk is tolerable cannot be solely based on a resolution of the facts.

The decision to take action in conditions of uncertainty bears little resemblance to the of empirically sort verifiable factual conclusions to which the substantial evidence test normally applied. Such decisions were to be intended unreviewable; they too must be scrutinized to ensure that the Secretary has acted reasonably and within the boundaries by set Congress. But a reviewing court must be mindful of the limited nature of its role. Yankee Nu- See Vermont clear Power Corp. NBDC, v. (1978). S. 519 must U. It recognize that the ultimate decision be based solely cannot on determinations of and fact, that those factual conclusions that have been reached are ones which the ill- courts are equipped to resolve on their own.

Under this standard of review, the decision to reduce the permissible exposure 1 ppm level to was within well Sec- retary’s authority. The of Appeals upheld Court the Secre- tary’s conclusions that benzene causes leukemia, blood dis- orders, and damage chromosomal even low levels, at that an level of 10 exposure ppm more is dangerous than one of 1 ppm, and that benefits will result from the proposed stand- ard. It did finding not set aside that his the number of lives that would be saved was subject to quantification. was reduction that conclusion his question did it Nor “feasible.” reason was Secretary’s decision circumstances, these language statutory with the conformance in full

able and adequately most which he “set that requiring avail best basis feasible, on the extent to the assures, impair material suffer will employee no that evidence, able employee if such even capacity or functional health of ment stand such by with dealt hazard to the exposure regular has (b) § 655 C. TJ. S. working life.” his period for the ard regu that conclude Secretary could record, the (5). On risk a definite pose level ppm1 above exposure lar but indeterminate some impairment in material resulting revealed Studies employees. number substantial possibly Expert exposure. to benzene attributable of deaths hundreds been had exposure level of no safe testified expert after ex declined the risk extent and shown incidence evidence direct some was There level. posure chromosomal disorders, blood nonmalignant leukemia, Moreover, below. ppm of 10 levels exposure damage an required existing evidence testified experts numerous hazardous. ppm1 above level inference testimony expert “well-reasoned —based stated We have evidence— empirical uncontradicted known on what first-hand when evidence’ 'substantial of itself may in Florida v. FPC unavailable.” ... question on the evidence Nothing 453, 464-465 S. Co., 404 U. Light Power & acting prevent purports Act a standard’s quantity toas information when definitive deficiency here, Where, as unavailable.26 is benefits examining prohibited say that to is not This among setting priorities process benefits costs relative bene costs consideration systematic substances, or hazardous Efforts standard-setting process. attempted is not fits help may generally, of reasons benefits, statements like quantify costs facilitate factors decisional consideration informed promote knowledge relates to the extent of the benefits rather than their I existence, see no reason to hold that the Secretary has exceeded his statutory authority.

B The plurality avoids this conclusion through reasoning that may charitably be described as obscure. According to the plurality, the definition of occupational safety and health standards as those “reasonably necessary appropriate provide safe or healthful . . . working conditions” requires the Secretary to show that it is “more likely than not” that the risk he seeks to regulate is a “significant” one. Ante, at 653. The plurality does not show how this requirement can be plausibly derived from the “reasonably necessary or ap- propriate” clause. Indeed, the plurality’s reasoning refuted by the Act’s language, structure, and legislative history, and it is foreclosed by every applicable guide to statutory con- struction. In short, the plurality’s standard is a fabrication bearing no connection with the acts or intentions of Congress.

At the outset, it is important to observe that “reasonably necessary or appropriate” clauses are routinely inserted in regulatory legislation, and in the past such clauses have uni- formly been interpreted general provisos that regulatory actions must bear a reasonable relation to those statutory purposes set forth in the statute’s substantive provisions. See, g., e. FCC v. National Citizens Committee Broad- casting, 436 775, U. S. 796-797 (1978); Mourning v. Family Publications Service, Inc., 411 U. S. 356, 369 (1973); Thorpe judicial review. See Dunlop v. Bachowski, 421 U. 560, S. 571-574 The Secretary indicates that he has attempted to quantify costs and ben efits in past. See 43 Reg. Fed. 54354, (1978) 54427-54431 (lead); id (cotton 27378-27379 dust). It is not necessary in present litigation say whether the Secretary must show a reasonable relation between costs and benefits. Discount- ing for the scientific uncertainty, the Secretary expressly reason- —and ably— found such a relation here. 268, Durham, 393 U. S. City Authority Housing v. (1969). today has never —until The Court 280-281 —inter- hav- clause as appropriate” necessary “reasonably a preted congres- specific a supersedes that content a substantive ing more is focused provision in a embodied directive sional principle, authority. This agency’s on an particularly the deter- understanding that course, the common reflects necessary” “reasonably are regulations of whether mination judgment the legislative reference only by made may be own, a court’s be based statute; it must not reflected taken should steps of what subjective view inevitably statutory goals. perceived promote “reasonably neces under the suggests that plurality The unless “unsafe” is not clause, workplace a sary” risk “significant” that a reviewing court to convince is able em particularly is approach Ante, 642. That is at issue. lan plain by the contradicted for it barrassing case, in this renders interpretation plurality’s The the Act. guage which, (b)(5), of § 655 sentence the first utterly superfluous Secretary to set the requires above, as noted will employee no . assures . most . adequately “which plural Indeed, health.” impairment of material suffer By the Act. out of sentence reads ity’s interpretation regulating standards the test makes plurality so doing, substantially physical agents harmful toxic substances op generally plainly for standards to the test identical — canon of odd And is an Congress intended. of what posite defini general vague ain insert that would construction that overcomes requirement clause threshold tional provision. standard-setting in a language placed specific demon statutory construction elementary principles most appro interpretation opposite precisely strate 380, 394-395 Inc., 417 U. S. v. g., FPC Texaco See, e. priate. 488- 332 U. S. Finanz-Korp., Uebersee Clark v. (1974); provided Congress could have short, existing action until regulatory take Secretary may not scientific proves evidence risk issue to be “signifi- 27 cant,” but it chose not to so. do

The plurality’s interpretation of the “reasonably necessary or appropriate” clause is also conclusively by leg- refuted islative history. While the standard-setting provision that the plurality ignores received legislative extensive attention, the definitional clause received none at all. An earlier version of the Act, see 8, supra, did n. not embody a feasibility clear constraint and was not restricted to toxic substances or to “material” impairments. The “reasonably necessary ap- propriate” clause was contained this prior version of the bill, as it was at all relevant times. In debating this ver- sion, Members of Congress repeatedly expressed concern that it would require a risk-free universe. See, g., ante, e. at 646- 649. The definitional clause was not mentioned at all, omission that be incomprehensible if Congress intended 27It compare useful to the Act with regulatory other statutes Congress which required has a showing of a relationship between costs and benefits or of an "unreasonable risk.” In some Congress statutes has expressly required analysis cost-benefit or a demonstration of some rea sonable relation between costs and benefits. See 33 (Flood U. S. C. 701a § Control Act of 1936); 42 U. S. C. (c)(2)(B) (1976 ed., §7545 Supp. II) (Clean Act); Air 33 U. S. C. (b)(4)(B) (1976 1314 § ed., II) Supp. (Clean Act). Water Congress others has imposed independent two requirements: that administrative action justified be “feasible” and by a balancing of costs and benefits, g., e. 43 (b) U. S. C. (1976 ed., § II) (Outer Supp. Continental Shelf Act); Lands (a) U. S. C. § (1976 (2)(D) II) ed., Supp. (Energy Policy Act). and Conservation This approach legislative demonstrates awareness of the difference between a feasibility constraint and a constraint based on weighing costs and ben infra, efits. See at 719-720. In still Congress others has authorized regu lation of risk,” “unreasonable a term which has been read some courts require balancing of costs and benefits. See, g., Aqua e. Slide ‘N’ Corp. Dive v. Consumer Safety Product Comm’n, (CA5 569 F. 2d 831 1978) (construing 15 U. S. C. (c)(2)(A) (Consumer §2058 Safety Product Act)); Forester v. Consumer Safety Product Comm’n, 182 U. S. App. D. C. (1977) 2dF. (construing 15 U. S. (s) (Child C. 1261 Toy Protection and Safety Act)). the risk quantify require clause by that demonstrate in order regulate sought he “significant.” on which history legislative portions only “rea to do ibid.-, nothing have relies, see

plurality which clause appropriate” necessary or sonably portions Those derived. finding” requirement “threshold version earlier toward directed criticisms consisted clause. already contained which definitional the statute subsequent amend by met turn, were criticisms, These will employee “no the strict application limited that ments feasi explicit an inserted substances, toxic to clause suffer” by the “impairment” word the modified constraint, and bility plu the for best at disingenuous It is “material.” adjective legislative statements isolated that suggest to rality subsequent were met that concerns expressing history, jus can finding, any “threshold” requiring not amendments necessary” “reasonably a into requirement such reading tify along.28 Act all inwas that clause “reasonably if perception that its on relies also plurality The it, there meaning ascribes it given not necessary” were clause toxic dealing those than other for “standards guidance be no For two Ante, at 640, 45. n. agents.” physical harmful materials “reasonably if the First, even force. argument is without reasons con if that even content, and independent have necessary” does clause any fairminded under it, it cannot describes plurality as the tent toxic sub (5) for (b) language express § supersede the reading agents. physical harmful stances “no applied bill version above, earlier an Second, as noted time, there At that all substances. language suffer” will employee “reason- argued that not be accordingly could it “gap,” no plurality content had clause appropriate” necessary or ably when be that reasoning must plurality’s light, the In this it. ascribes requirements (5) (b) the strict apply bill to Congress amended independent gained an clause substances, the definitional only toxic argu- surely this But standards. all comprehended in turn that meaning when It reasons head. their purposes congressional turns ment The plurality’s various structural arguments are also un- convincing. The fact finding “grave danger” re- quired ante, temporary standards, see 640, 45, at n. hardly implies the Secretary must permanent show for standards that it is probable more than not the substance to be regulated poses a “significant” risk. Nor reference is the to “toxic ante, materials,” at 643, any way informative. And the priority-setting ante, provision, 643-644, cannot plausibly be read to Secretary’s condition standard-setting authority on ability to meet the Court’s “threshold” requirement. plurality ignores applicable canons construction,

apparently because it finds existence inconvenient. their But as we stated quite recently, the inquiry statutory pur- into poses should be “informed by an regu- awareness lation is entitled to deference unless it can said be a supportable reasoned and interpretation of the Act.” Whirlpool Marshall, Corp. v. S. 11 (1980). U. Can honestly be said that the Secretary’s interpretation of the Act is “unreasoned” or “unsupportable”? And we stated case, same “safety legislation is to liberally construed to effectuate the congressional purpose.” Id., at 13. The plurality’s disregard of principles these gives credence the to frequently voiced criticism they that are only honored when the Court finds itself in agreement substantive with the agency action at issue.

In short, today’s decision represents a deci- usurpation of sionmaking authority that has been by properly exercised and belongs Congress and its representatives. authorized Congress singled special out toxic substances for regulation, it simultane- ously created a more (“reasonably lenient necessary”) for test standards generally, and once that more lenient test applicable, it some- superseded requirements how the strict for toxic That substances. rea- soning illogical is both Nor is circular. there basis for the ante, plurality’s suggestion, 649, see at n. original bill’s application “entirely to all standards was inadvertent.” support no in the statute’s construction has plurality’s

The structure, find- history. The threshold legislative language, own invention. plurality’s requires is ing plurality Congress, the acts or intentions relationship to bears no It views personal only reflecting understood and it can be allocation resources proper as to the plurality of the workplace. American safety

C conse- interested in the obviously more plurality intention discerning the decision than of its quences history of legislative language Congress. But since about the conjecture is no need for plain, there the Act are speculate, not for us to “It today’s decision. effects of its have altered Congress whether less act, much anticipated.” been this case specific events had stance know pretend I to Hill, do TVA S., 437 U. at 185. v. as a matter today is, erects plurality the test whether its dele- Congress and to that created preferable policy, uncertainty, scientific fraught with too the area is gates: for a court to policy, on considerations dependent too identi- require to it is desirable whether able determine administra- allowing an before risk “significant” fication of of the light But action. regulatory take agency tive out point necessary it is opinion, plurality tenor decision Congress’ one-sided, is not question that the regulation promulgate to authorize one. a reasonable issue here was benzene Secretary found this case unquantifiable definite albeit ppm posed at levels above disorders, nonmalignant blood damage, chromosomal risk of justify was sufficient evidence existing The and leukemia. it did but presented, a risk was that such conclusion Discounting risk. of that rough quantification even permit gave Secretary uncertainties, scientific for the various “careful to the consideration question of whether sub- th[e] stantial costs” of the justified standard “are in light hazards exposure benzene,” and concluded that “these costs are necessary in order to effectuate the statutory pur- pose . . . and to adequately protect employees from the hazards of exposure to benzene.” 43 Reg. Fed. 5941 (1978).

In these circumstances it seems clear that the Secretary found a risk “significant” in the sense that the word is normally used. There was some direct evidence of chro- mosomal damage, nonmalignant blood disorders, and leukemia at exposures at or near ppm10 and below. In addition, ex- pert after expert testified that the recorded effects of benzene higher at justified levels an inference that an ex- posure level above ppm was dangerous. plurality’s extraordinarily searching scrutiny of this factual re- record veals no basis for a conclusion that quantification is, on the basis “the best available evidence,” possible at the present If time. the Secretary decided to wait until infor- definitive mation was available, American workers would be subjected for the indefinite future to a possibly substantial risk of ben- zene-induced leukemia and other illnesses. It is unsurpris- ing, least to me, that he concluded that the statute author- ized him to take regulatory action now.

Under these circumstances, the plurality’s requirement of identification of a “significant” risk will have one of two con- sequences. If the plurality means require realistically to “quantify” the risk in order to satisfy a court it is “significant,” the record shows that the plurality means to require him to do the impossible. But regulatory inaction has very significant costs of its own. The adoption of such a test would subject American workers ato continuing risk of cancer and other serious diseases; it would disable Secretary from regulating wide variety of carcinogens for which quantification simply cannot be undertaken at present time. not signs today’s decision does encouraging

There are that My Brother concludes that Powell extend that far.29 action taking regulatory from Secretary prevented is not accomplished by cannot be quantification “when reasonable ante, plurality The any known methods.” See at 666. Secretary prohibit indicates not also it would ben- of the quantification when safety standards promulgating ante, 63. The 656-657, is at n. impossible. efits See to make attempt Secretary allow the to might thus Court by a carcino- imposed risk very rough quantification his find- give deference genic substance, considerable per- would If the Court significant. so, risk ing regulation the same Secretary precisely to promulgate mit the carcinogen on a not relied if he had involved these cases and the of the evidence a review but undertaken “policy,” determine, in the agency for the “to suggests that it plurality risk,” 'significant’ be a instance, what it considers to first interpreting the assumptions in agency free conservative “is to use not would my Ante, Moreover, Brother Powell 656. data. . . .” at (opinion con every Ante, at 666 case.” require “quantification of risk in opinion, I read his judgment). As concurring in curring part Secretary promulgate permitted the would Mr. Justice Powell have carefully Secretary provided a more had here if at issue justified the at issue that the risk explanation conclusion of his reasoned Powell admittedly significant Mr. Justice the benzene standard. costs of subject relatively deferen suggests that such a conclusion also 670-671, Ante, n. 8. tial review. my approach and that differences between respect, In this agreed are narrow. We may comparatively Mr. Justice Powell interpretation regard critical to a fairminded that I propositions two subject may regulate that are Secretary risks (1) of the Act: evidence”; and available the “best quantification basis of on the regu- risk merits particular health Secretary’s judgment (2) scrutiny. encouraging judicial It subject latory to limited action is propositions. accept basic Court these Members that at least five my text, however, disagree I Brother For reasons stated that the appropriate to in these cases it is hold conclusion that Powell’s relationship between a reasonable show requires Act and benefits. costs

expert testimony and concluded, on the basis of conservative assumptions, the risk addressed significant is a one. Any other interpretation of the plurality’s approach al- would low a court to displace the agency’s judgment with its own subjective conception of “significance,” a duty to be per- formed without statutory guidance. The consequences of this approach second be hardly

disastrous; indeed, it differs from my principally own in its assessment of the basis for the Secretary’s decision in these cases. It is objectionable, however, for three reasons. First, the requirement of identification of “significant” a risk sim- ply has no relationship to the statute that today Court purports to Second, construe. if the “threshold finding” re- quirement means only that the Secretary must find “that there is a need for such ante, at 643, n. standard,” 48, requirement plainly by satisfied the Secretary’s express statement that the standard’s costs “are necessary in order to effectuate the statutory purpose . .. and to adequately pro- tect employees from the hazards of exposure to benzene.” 43 Fed. Reg. Third, the record amply demon- strates in light of existing scientific knowledge, no pur- pose would be served requiring the Secretary to steps take to quantify the risk exposure to benzene at low levels. Any such quantification would be based on not scientific “knowledge” as that term is normally understood, but con- siderations policy. For carcinogens like benzene, the as- sumptions on which a dose-response curve must be based are arbitrary. necessarily require To quantitative showing therefore, a “significant” risk, would either paralyze the Sec- retary into inaction or force him to deceive public by act- ing on the basis of assumptions that must be considered too speculative to support any realistic assessment of the relevant risk. See McGarity, Substantive and Procedural Discretion in Administrative Resolution of Science Policy Questions: Regulating Carcinogens OSHA, in EPA and 729, 67 Geo. L. J. 806 (1979). It is encouraging that the Court appears willing require quantification when it fairly is not possible. 656-657, ante,

See and n. 63. Though it is to see difficult how a future Congress could more explicit on the matter than was Congress passed inAct it is important to remember that today’s decision subject to legislative reversal. Congress may continue to believe the Secretary should not pre- vented from protecting American workers from cancer and other fatal diseases until scientific evidence progressed has to a point where he can convince a federal court that the risk is “significant.” Today’s objectionable decision is not be- cause it is final, but places because it of legis- burden lative inertia on the beneficiaries of safety and health legislation in question in these cases. By allocating the bur- den in this fashion, the Court requires the American worker to return to political arena and to victory win a that he won once before I 1970. am unable to any justi- discern fication for that result.

D Since the plurality’s construction of the “reasonably neces- sary or appropriate” clause is I turn a brief unsupportable, discussion of the arguments other respondents offer in support judgment of the below.

First, respondents characterize the Act as a pragmatic statute designed to balance the benefits a safety and health regulation against its Respondents costs. observe that statute speaks terms of protection relative by providing that safety must assured “so far as possible,” 29 U. S. C. § 651 (b), by stating that the “no impairment” material requirement is to be imposed only “to the feasible.” extent RehNQüist obscurity Finding my word “feasible,” Brother nondelegation invokes the doctrine, which was last used to invalidate an Act of Congress in 1935. L. Poultry A. Corp. A. Schecter States, v. United Rehnquist 295 U. S. my While Brother eloquently argues there place remains a for such a jurisprudence, doctrine in our I am frankly puzzled why as to thought the issue to be of relevance Respondents contend that the term “feasible” should be read to require consideration the economic burden of a stand- not merely ard. its technological achievability. I do not understand Secretary to disagree. respondents But pre- sent no argument the expenditure required by ben- zene standard is not feasible in that respect.

concluded on the basis substantial evidence that the costs of the standard would be readily by absorbed the 20 affected industries. One need not define feasibility requirement precision in order to conclude that the benzene standard is “feasible” in the sense that it will not harm materially financial condition of regulated industries.

Respondents suggest that requirement the feasibility should be understood merely to refer to a expense, standard’s but also to mandate a finding that the benefits of an occupa- tional safety and health standard bear a reasonable relation nondelegation here. The designed doctrine is to assure that the fun most damental by decisions will be Congress, made representatives the elected people, by rather than- administrators. Some minimal definiteness is therefore required in Congress order for delegate authority its agencies. administrative Congress has been sufficiently definite here. The word “feasible” has a reasonably plain meaning, interpretation and its can be informed other contexts in which Congress has used it. See supra. n. Since the placed term in the same sentence with the employee “no will suffer” language, is clear “feasible” means technologically and eco- nomically achievable. Under Act, the Secretary is afforded consid- erably more guidance than are other acting administrators under different regulatory short, statutes. Congress has made “the policy critical ante, cases, decisions” in these see 687 (Rehnquist, at J., concurring in ,judgment). plurality’s The apparent suggestion, ante, see 646, the nondele- gation might doctrine be violated if the Secretary permitted were regulate nonquantifiable definite but risks plainly wrong. Such a quite statute would be definite and would thus raise no constitutional question Poultry. under Schechter Moreover, Congress could rationally decide that it would be require better to industry to bear “feasible” costs subject than to American workers an indeterminate risk of cancer and other fatal diseases. structure, language, statute’s I believe that costs.

to its In its position. respondents’ history foreclose legislative and of capable if it is activity is “feasible” meaning an ordinary Web- See outweigh its costs. achievement, its if benefits not More- Dictionary 831 International Third New ster’s (5) (b) 655§ render would interpretation over, respondents’ “feasible” the term reading into by internally inconsistent author- language express irreconcilable requirement em- “no assuring standards set Secretary to izing Respondents’ . .” . impairment. material suffer will ployee As hortatory. merely language render that position to at referred analysis is cost-benefit no above, noted omission history, legislative or its in the statute point explicit light inadvertent deemed cannot which legis- regulatory into other inserted requirements cost-benefit re- feasibility history of legislative Finally, the lation.31 sole Congress’ demonstrates supra, 8, n. see quirement, technolog- economically standards was that concern prevent was to intent legislative ically achievable. condition financial harming materially from Secretary impairment. risks eliminate order to industries regulated taking preclude intend did not Congress industry threat such no here, where, action regulatory posed.32 supra. n. See through balancing is evident cost-benefit antipathy toward Congress’ example: For history Act. legislative out meeting the standards employers in by incurred will be costs “The view, my are, bill under safety be established health in- we, as Whether business. doing necessary costs reasonable economics, we simple by humanity or simple dividuals, motivated are unhealthy upon an unsafe dependent to be profits permit longer no can (Sen. 1150-1151 Leg. Hist. (1970), Cong. Rec worksite.” Eagieton). Yarborough stated:

Similarly, Senator cost lives, some indifference people’s talking about are “We who men women assuring the talking about areWe accountants. In order to decide these cases, however, it is not necessary to resolve the question whether the term “feasible” con- may template some balancing of the costs and benefits of regula- tory action.33 into Taking account the uncertainties in exist- ing knowledge, the Secretary an made express finding the hazards of benzene exposure were sufficient to justify the regulation’s costs. 43 Fed. Reg. 5941 (1978). Any require- ment to balance costs and benefits cannot be read to invali- date this wholly rational conclusion. A contrary result, forc- ing the Secretary to wait for quantitative data that may not be available in the future, foreseeable would run directly counter to the protective purposes of the Act.34 work plants in our and factories they thát gowill home after day’s work with their bodies intact. We talking are about assuring our American workers who deadly wo[r]k chemicals they that when have accumulated a few year’s seniority they will not have lung accumulated congestion poison bodies, their or something that will strike them down they before reach retirement age.” Cong. Rec., at 37625, Leg. Hist. 510. 33Nor need I discuss the possibility, by raised counsel for the federal parties in oral argument, that a regulate decision to a substance posing a negligible threat to health safety could itself challenged arbitrary as capricious under the Administrative Procedure Act. See Tr. of Oral Arg. 23. 34Respondents rely also on the statutory requirement that the Secre tary may act only prevent "material” impairment. They contend the standard promulgated here does not fall within that category because the risk is so low. interpretation This derives no support from the statute or legislative its history. The statute itself states that standards should ensure that no employee will suffer impair “material ment,” not material risk impairment. The language is consistent with legislative history. early version the Act, the word “impairment” was modified “any” rather than “material.” See supra. n. The feasibility and materiality requirements were added simultaneously part of an effort qualify original language authorizing the Secretary to ensure that “no em- ployee will suffer any impairment of health or functional capacity, or diminished life expectancy.” Senator Dominick was concerned that

Finally, respondents suggest broadly Secretary that the did not fulfill his statutory responsibility the to act on basis of “research, demonstrations, experiments,” and to consider “the field, latest available scientific data in feasibility the the of standards, experience gained under this and other Here, safety health and laws.” (b)(5). 29 U. §655 S. C. contend, they the Secretary solely based his decision arguments.” “views and Respondents Brief for American Petroleum Institute Secretary et al. 52. I disagree. The compiled an extensive composed record of over 50 volumes exhibits. Most of reported those exhibits are results research and representing demonstrations “the latest avail- able scientific Secretary data.” The offered careful discus- sion of these data in the accompanying perma- statement nent standard. His ultimate conclusions grounded were findings extensive Where, here, fact. there gaps are existing knowledge, Secretary’s necessarily decision must be based on of policy considerations on empirically as well as verifiable facts.

In passing Occupational Safety and Health of 1970, Act Congress was aware that authorizing Secretary it was regulate to areas of scientific uncertainty. But intended require stringent regulation to even when definitive informa- tion was unavailable. In reducing permissible level benzene, Secretary applied proper to stand- legal His supported ards. determinations are by substantial evi- phrase “any” impairment require Secretary prevent to insect Cong. (1970), Leg. bites. Rec. 36522 Hist. 345. respondents’ pose

The construction would an enormous obstacle to regulate (b)(5). probability efforts toxic substances under §655 contracting quite cancer will in respect most contexts be small with any particular employee. If the statute were read to authorize only Secretary employee to assure that act “no will suffer material impairment,” risk would be from regulating disabled poses which a small risk respect any particular substances will employee but which nonetheless result in the death of numerous employee pool. members of the The Secretary’s

denee. one, then, decision was which the governing legislation authorized him make.35

IV years increasing recognition recent there has been products the harmful technological development may have effects severity predicted whose incidence and cannot certainty. The responsibility products to such has regulate to fallen an en- agencies. administrative Their task is not Frequently viable one. no causal link can be established clear regulated between the substance and the harm to be averted. harm Risks of are often inaction considera- uncertain, but has ble of its whether to agency own. The must decide take costs risks or wait regulatory against possibly action substantial judg- until more becomes definitive information available —a Appeals Secretary’s finding that Although accepted the Court of the leukemia, the dermal with benzene could it set aside dermal contact cause Secretary’s perform experi failure to an contact standard because by industry this The failure to conduct ment recommended witness. requirement test, according court, statutory to the violated the Secretary “the latest act of “the best evidence” on basis available available scientific data in the field.” agency, respondents presented substantial hearings before the no

In the skin, through challenge position be absorbed to the that benzene could support position. Both evidence in the record to and there was absorption. In circum- had such these animal and human studies found obligated stances, undertake additional studies was not informative. studies would be simply because a witness testified that such requirement paralyze would the standard- imposition The of such Secretary’s the basis process. mandate is act on setting The may evidence, become available in the not which evidence “available” future. standard, Appeals also Court setting the dermal contact aside In quanti- Secretary had not shown that conclusion on its

relied above As the discussion from the standard. result would fiable benefits holding. legal standards so incorrect indicates, applied the court solely based deter- nature cannot be by very which its ment of fact.36 minations turn, under made on the have been delegations,

Those ensure that available to judicial would be standing that review by substantial are supported determinations agency’s limits set its do not and that exceed evidence actions Act, Con Health Occupational Safety Congress. carry out the courts confidence expressed gress plurality in these cases the But responsibility. important “threshold authority. plurality’s its has far exceeded Act and be found nowhere to finding” requirement “The fundamental purposes. its basic is antithetical . are . . Congress resolved appropriately policy questions courts under in the federal re-examination subject ' Yankee Vermont agency action.” review judicial guise (emphasis at 558 NRDC, S., Nuclear Power Corp. v. U. en decision of the no less true Surely this is original). pro decision worker than American safety sure See ibid. power. with nuclear ceed plainly sois by plurality taken approach Because role, I am institutional proper the Court’s irreconcilable likeli- all of time. stand test it will not certain that extreme as an regarded be come to will hood, today’s decision Members as the that, scheme regulatory ato reaction harsh burden unduly imposed an perceived plurality it, not “does as the Constitution But industries. regulated Newv. Statics,” Lochner Social Spencer’s Mr. Herbert enact so the dissenting), (Holmes, J., (1905) York, 45, 198 U. S. does action federal administrative scrutinize responsibility between balance its own strike Court to authorize and the Determina Science Acceptable Risk: Lowrance, Of See W. Liberty, Integrity Stewart, Paradoxes (1976); Safety tion Quality Judicial Environmental Fraternity: Nature The Collective 469-472 L. Action, 7 Environ. Administrative Review of occupational costs and safety benefits of I standards. am confident approach that the taken the plurality like today, Lochner that in eventually itself, will and that abandoned, representative government branches of will again once allowed to level safety determine the protec- and health tion to be accorded to the American worker.

Case Details

Case Name: Industrial Union Dept., AFL-CIO v. American Petroleum Institute

Court Name: Supreme Court of the United States

Date Published: Jul 2, 1980

Citation: 448 U.S. 607

Docket Number: 78-911

Court Abbreviation: SCOTUS