concurring in part and concurring in the judgment.
I join Parts I, II, III-A, III-B, III-C, and III-E of the plurality opinion.1 The Occupational Safety and Health Administration relied in large part on its “carcinogen policy”— which had not been adopted formally — in promulgating the benzene exposure and dermal contact regulation at issue in these cases.2 For the reasons stated by the plurality, I agree that §§ 6 (b)(5) and 3(8) of the Occupational Safety and Health Act of 1970, 29 U. S. C. §§ 655 (b)(5) and 652 (8), must be read together. They require OSHA to make a threshold finding that proposed occupational health standards are reasonably necessary to provide safe workplaces. When OSHA acts to reduce existing national consensus standards, *665therefore, it must find that (i) currently permissible exposure levels create a significant risk of material health impairment; and (ii) a reduction of those levels would significantly reduce the hazard.
Although I would not rule out the possibility that the necessary findings could rest in part on generic policies properly adopted by OSHA, see McGarity, Substantive and Procedural Discretion in Administrative Resolution of Science Policy Questions: Regulating Carcinogens in EPA and OSHA, 67 Geo. L. J. 729, 754-759 (1979), no properly supported agency policies are before us in these cases.3 I therefore agree with the plurality that the regulation is invalid to the extent it rests upon the assumption that exposure to known carcinogens always should be reduced to a level proved to be safe or, if no such level is found, to the lowest level that the affected industry can achieve with available technology.
I
If the disputed regulation were based exclusively on this “carcinogen policy,” I also would agree that we need not consider whether the Act requires OSHA to determine that the benefits of a proposed standard are reasonably related to the costs of compliance. Ante, at 615. As the Court of Appeals for the Fifth Circuit recognized, however, OSHA takes the “fall-back position” that its regulation is justified by specific findings based upon the voluminous evidentiary record compiled in this case. American Petroleum Institute v. OSHA, 581 F. 2d 493, 503. OSHA found, for example, that the num*666ber of cancers prevented by reducing permissible exposure levels from 10 ppm to 1 ppm “may be appreciable,” that “the benefits of the proposed standard are likely to be appreciable,” and that the “substantial costs [of the new standard] are justified in light of the hazards.” 43 Fed. Reg. 5940-5941 (1978). Thus, OSHA found — at least generally — that the hazards of benzene exposure at currently permissible levels are serious enough to justify an expenditure of hundreds of millions of dollars. For me, that finding necessarily subsumes the conclusion that the health risk is “significant.” If OSHA’s conclusion is supported by substantial evidence, the threshold requirement discussed in the plurality opinion would be satisfied.
As I read its opinion, the plurality does not consider whether the agency’s findings are supported by substantial evidence. The Court of Appeals found them insufficient because OSHA failed “to estimate the extent of expected benefits. . . .” 581 F. 2d, at 504. That court apparently would have required OSHA to supply a specific numerical estimate of benefits derived through mathematical techniques for “risk quantification” or “cost-effectiveness analysis.” Id., at 504, n. 23; see id., at 504-505. I do not agree with the Court of Appeals’ conclusion that the statute requires quantification of risk in every case.
The statutory preference for the “best available evidence,” 29 U. S. C. § 655 (b)(5), implies that OSHA must use the 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 reasonable quantification cannot be accomplished by any known methods. See post, at 693 (Marshall, J., dissenting). In this litigation, OSHA found that “it is impossible to precisely quantify the anticipated benefits. . . .” 43 Fed. Reg. 5941 (1978). If this finding is supported by substantial evidence, the statute does not prevent the Secretary from finding a significant health hazard on the *667basis of the weight of expert testimony and opinion. I do not understand the plurality to hold otherwise. See ante, at 662.
For the foregoing reasons. I would not hold that “OSHA did not even attempt to carry its burden of proof” on the threshold question whether exposure to benzene at 10 ppm presents a significant risk to human health. Ante, at 653. In my view, the question is whether OSHA successfully carried its burden on the basis of record evidence. That question in turn reduces to two principal issues. First, is there substantial evidence supporting OSHA’s determination that available quantification techniques are too imprecise to permit a reasonable numerical estimate of risks? If not, then OSHA has failed to show that its regulation rests on the “best available -evidence.” Second, is OSHA’s finding of significant risks at current exposure levels supported by substantial evidence? If not, then OSHA has failed to show that the new regulation is reasonably necessary to provide safe and healthful workplaces.
II
Although I regard the question as close, I do not disagree with the plurality’s view that OSHA has failed, on this record, to carry its burden of proof on the threshold issues summarized above. But even if one assumes that OSHA properly met this burden, see post, at 697-701, 713-714 (Marshall, J., dissenting), I conclude that the statute also requires the agency to determine that the economic effects of its standard bear a reasonable relationship to the expected benefits. An occupational health standard is neither “reasonably necessary” nor “feasible,” as required by statute, if it calls for expenditures wholly disproportionate to the expected health and safety benefits.
OSHA contends that § 6 (b) (5) not only permits but actually requires it to promulgate standards that reduce health risks without regard to economic effects, unless those effects *668would cause widespread dislocation throughout an entire industry.4 Under the threshold test adopted by the plurality today, this authority will exist only with respect to “significant” risks. But the plurality does not reject OSHA’s claim that it must reduce such risks without considering economic consequences less serious than massive dislocation. In 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 *669It is simply unreasonable to believe that Congress intended OSHA to pursue the desirable goal of risk-free workplaces to the extent that the economic viability of particular industries — or significant segments thereof — is threatened. As the plurality observes, OSHA itself has not chosen to carry out such a self-defeating policy in all instances. Ante, at 650. If it did, OSHA regulations would impair the ability of American industries to compete effectively with foreign businesses and to provide employment for American workers.6
I therefore would not lightly assume that Congress intended OSHA to require reduction of health risks found to be significant whenever it also finds that the affected indus*670try can bear the costs. See n. 4, supra. Perhaps more significantly, however, OSHA’s interpretation of §6 (b)(5) would force it to regulate in a manner inconsistent with the important health and safety purposes of the legislation we construe today. Thousands of toxic substances present risks that fairly could be characterized as “significant.” Cf. ante, at 645, n. 51. Even if OSHA succeeded in selecting the gravest risks for earliest regulation, a standard-setting process that ignored economic'considerations would result in a serious misallocation of resources and a lower effective level of safety than could be achieved under standards set with reference to the comparative benefits available at a lower cost.7 I would not attribute such an irrational intention to Congress.
In these cases, OSHA did find that the “substantial costs” of the benzene regulations are justified. See supra, at 665-666. But the record before us contains neither adequate documentation of this conclusion, nor any evidence that OSHA weighed the relevant considerations. The agency simply announced its finding of cost-justification without explaining the method by which it determines that the benefits justify the costs and their economic effects. No rational system of regulation can permit its administrators to make policy judgments without explaining how their decisions effectuate the purposes of the governing law. and nothing in the statute authorizes such laxity in these cases.8 Since neither the air*671borne concentration standard nor the dermal contact standard for exposure to benzene satisfies the requirements of the governing statute, I join the Court's judgment affirming the judgment of the Court of Appeals.
These portions of the plurality opinion primarily address OSHA’s special carcinogen policy, rather than OSHA’s argument that it also made evidentiary findings. I do not necessarily agree with every observation in the plurality opinion concerning the presence or absence of such findings. I also express no view on the question whether a different interpretation of the statute would violate the nondelegation doctrine of A. L. A. Schechter Poultry Corp. v. United States, 295 U. S. 495 (1935), and Panama Refining Co. v. Ryan, 293 U. S. 388 (1935). See post, at 672-687 (Rehnquist, J., concurring in judgment).
The Secretary of Labor promulgated the relevant standard pursuant to his statutory authority. Since OSHA is the agency responsible for developing such regulations under the Secretary’s direction, this opinion refers to “OSHA” or “the agency” as the decisionmaker most directly concerned.
OSHA has adopted a formal policy for regulating carcinogens effective April 21, 1980. 45 Fed. Reg. 5282 (1980) (to be codified at 29 CFR, Part 1990). But no such policy was in effect when the agency promulgated its benzene regulation. Moreover, neither the factual determinations nor the administrative judgments upon which the policy rests are supported adequately on this record alone. Accordingly, we have no occasion to consider the extent to which valid agency policies may supply a basis for a finding that health risks exist in particular cases.
OSHA argues that § 6 (b) (5) requires it to promulgate standards that are “feasible” only in the sense that they are “capable of achievement”; that 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 affected industry. See, e. g., American Federation of Labor v. Brennan, 530 F. 2d 109, 123 (CA3 1975). In this case, OSHA simply asked a consulting firm to ascertain the costs of complying with a 1 ppm standard. See ante, at 621. OSHA then concluded that “the economic impact of [compliance] will not . . . threaten the financial welfare of the affected firms or the. general economy.” 43 Fed. Reg. 5939 (1978). The cost of complying with a standard may be “bearable” and still not reasonably related to the benefits expected. A manufacturing company, for example, may have financial resources that enable it to pay the OSHA-ordered costs. But expenditures for unproductive purposes may limit seriously its financial ability to remain competitive and provide jobs.
I will not repeat the detailed summary of the legislative history contained in the plurality opinion. Ante, at 646-652. Many of the considerations that the plurality relies upon to show Congress’ concern with significant harms persuade me that Congress did not intend OSHA to reduce each significant hazard without regard to economic consequences. Senator Williams, a sponsor of the legislation, stated: “Our bill is fair and reasonable. It is a good-faith effort to balance the need of workers to have a sa[f]e and healthy work environment against the requirement of industry to function without undue interference.” 116 Cong. Rec. 37342 (1970), Legislative History of the Occupational Safety and Health Act of 1970 (Committee Print compiled for the Senate Committee on Labor and Public Welfare), p. 435 (1971). There could be no such “balance” if OSHA were *669authorized to impose standards without regard to economic consequences short of serious dislocation.
Senator Dominick described a preliminary version of § 6 (b) (5) as follows:
“What we were trying to do in the bill . . . was to say that when we are dealing with toxic agents or physical agents, we ought to take such steps as are feasible and practical to provide an atmosphere within which a person’s health or safety would not be affected. Unfortunately, we had language providing that anyone [sic] would be assured that no one would have a hazard. . . .
“It was an unrealistic standard. . . .” 116 Cong. Rec. 37622 (1970), Legislative History, supra, at 502 (emphasis added).
Senator Dominick’s objection to the “unrealistic” standard of the forerunner of § 6 (b) (5) does not imply that he thought § 3 (8) of the Act lacked substantive content. See post, at 710-711 (Marshall, J., dissenting) . The Senator hardly would have proposed that § 6 (b) (5) be deleted entirely, see ante, at 647, if he had not thought that other sections of the Act required health regulations that were reasonable and practical.
Congress has assigned OSHA an extremely difficult and complex task, and the guidance afforded OSHA is considerably less than clear. The agency’s primary responsibility, reflected in its title, is to minimize health and safety risks in the workplace. Yet the economic health of our highly industrialized society requires a high rate of employment and an adequate response to increasingly vigorous foreign competition. There can be little doubt that Congress intended OSHA to balance reasonably the societal interest in health and safety with the often conflicting goal of maintaining a strong national economy.
For example, OSHA’s reading of § 6 (b)(5) could force the depletion of an industry’s resources in an effort to reduce a single risk by some speculative amount, even though other significant risks remain unregulated.
The decision that costs justify benefits is largely a policy judgment delegated to OSHA by Congress. When a court reviews such judgments under the "substantial evidence” standard mandated by 29 U. S. C. § 655 (f), the court must determine whether the responsible agency has earefulLly] identifi[ed] . . the reasons win ¡ir] chooses to follow one course rather than another” as the most reasonable method of effectuating the purposes of the applicable law. Industrial Union Dept. v. Hodgson, 162 U. S. App. D. C. 331, 339-340, 499 F. 2d 467, 475-476 (1974). Since OSHA failed to identify its reasons in these cases, I express no *671opinion as to the standard of review that may be appropriate in other situations.