CHEVRON U.S. A. INC. V. ECHAZABAL (00-1406) 536 U.S. 73 (2002)
226 F.3d 1063, reversed and remanded.
[ Souter ]
HTML version
PDF version
HTML version
PDF version

Opinion of the Court

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.


No. 00—1406



[June 10, 2002]

Justice Souter delivered the opinion of the Court.

A regulation of the Equal Employment Opportunity Commission authorizes refusal to hire an individual because his performance on the job would endanger his own health, owing to a disability. The question in this case is whether the Americans with Disabilities Act of 1990, 104 Stat. 328, 42 U.S.C. § 12101 et seq. (1994 ed. and Supp. V), permits the regulation.1 We hold that it does.


Beginning in 1972, respondent Mario Echazabal worked for independent contractors at an oil refinery owned by petitioner Chevron U.S. A. Inc. Twice he applied for a job directly with Chevron, which offered to hire him if he could pass the company’s physical examination. See 42 U.S.C. § 12112(d)(3) (1994 ed.). Each time, the exam showed liver abnormality or damage, the cause eventually being identified as Hepatitis C, which Chevron’s doctors said would be aggravated by continued exposure to toxins at Chevron’s refinery. In each instance, the company withdrew the offer, and the second time it asked the contractor employing Echazabal either to reassign him to a job without exposure to harmful chemicals or to remove him from the refinery altogether. The contractor laid him off in early 1996.

Echazabal filed suit, ultimately removed to federal court, claiming, among other things, that Chevron violated the Americans With Disabilities Act in refusing to hire him, or even to let him continue working in the plant, because of a disability, his liver condition.2 Chevron defended under a regulation of the Equal Employment Opportunity Commission permitting the defense that a worker’s disability on the job would pose a “direct threat” to his health, see 29 CFR § 1630.15(b)(2) (2001). Although two medical witnesses disputed Chevron’s judgment that Echazabal’s liver function was impaired and subject to further damage under the job conditions in the refinery, the District Court granted summary judgment for Chevron. It held that Echazabal raised no genuine issue of material fact as to whether the company acted reasonably in relying on its own doctors’ medical advice, regardless of its accuracy.

On appeal, the Ninth Circuit asked for briefs on a threshold question not raised before, whether the EEOC’s regulation recognizing a threat-to-self defense, ibid., exceeded the scope of permissible rulemaking under the ADA. 226 F.3d 1063, 1066, n. 3 (2000). The Circuit held that it did and reversed the summary judgment. The court rested its position on the text of the ADA itself in explicitly recognizing an employer’s right to adopt an employment qualification barring anyone whose disability would place others in the workplace at risk, while saying nothing about threats to the disabled employee himself. The majority opinion reasoned that “by specifying only threats to ‘other individuals in the workplace,’ the statute makes it clear that threats to other persons–including the disabled individual himself–are not included within the scope of the [direct threat] defense,” id., at 1066—1067, and it indicated that any such regulation would unreasonably conflict with congressional policy against paternalism in the workplace, id., at 1067—1070. The court went on to reject Chevron’s further argument that Echazabal was not “ ‘otherwise qualified’ ” to perform the job, holding that the ability to perform a job without risk to one’s health or safety is not an “ ‘essential function’ ” of the job. Id., at 1070.

The decision conflicted with one from the Eleventh Circuit, Moses v. American Nonwovens, Inc., 97 F.3d 446, 447 (1996), and raised tension with the Seventh Circuit case of Koshinski v. Decatur Foundry, Inc., 177 F.3d 599, 603 (1999). We granted certiorari, 534 U.S. 991 (2001), and now reverse.


Section 102 of the Americans with Disabilities Act of 1990, 104 Stat. 328, 42 U.S.C. § 12101 et seq., prohibits “discriminat[ion] against a qualified individual with a disability because of the disability … in regard to” a number of actions by an employer, including “hiring.” 42 U.S.C. § 12112(a). The statutory definition of “discriminat[ion]” covers a number of things an employer might do to block a disabled person from advancing in the workplace, such as “using qualification standards … that screen out or tend to screen out an individual with a disability.” §12112(b)(6). By that same definition, ibid., as well as by separate provision, §12113(a), the Act creates an affirmative defense for action under a qualification standard “shown to be job-related for the position in question and … consistent with business necessity.” Such a standard may include “a requirement that an individual shall not pose a direct threat to the health or safety of other individuals in the workplace,” §12113(b), if the individual cannot perform the job safely with reasonable accommodation, §12113(a). By regulation, the EEOC carries the defense one step further, in allowing an employer to screen out a potential worker with a disability not only for risks that he would pose to others in the workplace but for risks on the job to his own health or safety as well: “The term ‘qualification standard’ may include a requirement that an individual shall not pose a direct threat to the health or safety of the individual or others in the workplace.” 29 CFR § 1630.15(b)(2) (2001).

Chevron relies on the regulation here, since it says a job in the refinery would pose a “direct threat” to Echazabal’s health. In seeking deference to the agency, it argues that nothing in the statute unambiguously precludes such a defense, while the regulation was adopted under authority explicitly delegated by Congress, 42 U.S.C. § 12116 and after notice-and-comment rulemaking. See United States v. Mead Corp., 533 U.S. 218, 227 (2001); Chevron U.S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842—844 (1984). Echazabal, on the contrary, argues that as a matter of law the statute precludes the regulation, which he claims would be an unreasonable interpretation even if the agency had leeway to go beyond the literal text.


As for the textual bar to any agency action as a matter of law, Echazabal says that Chevron loses on the threshold question whether the statute leaves a gap for the EEOC to fill. See id., at 843—844. Echazabal recognizes the generality of the language providing for a defense when a plaintiff is screened out by “qualification standards” that are “job-related and consistent with business necessity” (and reasonable accommodation would not cure the difficulty posed by employment). 42 U.S.C. § 12113(a). Without more, those provisions would allow an employer to turn away someone whose work would pose a serious risk to himself. That possibility is said to be eliminated, however, by the further specification that “ ‘qualification standards’ may include a requirement that an individual shall not pose a direct threat to the health or safety of other individuals in the workplace.” §12113(b); see also §12111(3) (defining “direct threat” in terms of risk to others). Echazabal contrasts this provision with an EEOC regulation under the Rehabilitation Act of 1973, 87 Stat. 357, as amended, 29 U.S.C. § 701 et seq., antedating the ADA, which recognized an employer’s right to consider threats both to other workers and to the threatening employee himself. Because the ADA defense provision recognizes threats only if they extend to another, Echazabal reads the statute to imply as a matter of law that threats to the worker himself cannot count.

The argument follows the reliance of the Ninth Circuit majority on the interpretive canon, expressio unius exclusio alterius, “expressing one item of [an] associated group or series excludes another left unmentioned.” United States v. Vonn, 535 U.S. ___ , ___ , (2002) (slip op., at 8). The rule is fine when it applies, but this case joins some others in showing when it does not. See, e.g., id., at ___ (slip op., at 9); United Dominion Industries, Inc. v. United States, 532 U.S. 822, 836 (2001); Pauley v. BethEnergy Mines, Inc., 501 U.S. 680, 703 (1991).

The first strike against the expression-exclusion rule here is right in the text that Echazabal quotes. Congress included the harm-to-others provision as an example of legitimate qualifications that are “job-related and consistent with business necessity.” These are spacious defensive categories, which seem to give an agency (or in the absence of agency action, a court) a good deal of discretion in setting the limits of permissible qualification standards. That discretion is confirmed, if not magnified, by the provision that “qualification standards” falling within the limits of job relation and business necessity “may include” a veto on those who would directly threaten others in the workplace. Far from supporting Echazabal’s position, the expansive phrasing of “may include” points directly away from the sort of exclusive specification he claims. United States v. New York Telephone Co., 434 U.S. 159, 169 (1977); Federal Land Bank of St. Paul v. Bismarck Lumber Co., 314 U.S. 95, 100 (1941).3

Just as statutory language suggesting exclusiveness is missing, so is that essential extrastatutory ingredient of an expression-exclusion demonstration, the series of terms from which an omission bespeaks a negative implication. The canon depends on identifying a series of two or more terms or things that should be understood to go hand in hand, which are abridged in circumstances supporting a sensible inference that the term left out must have been meant to be excluded. E. Crawford, Construction of Statutes 337 (1940) (expressio unius ‘properly applies only when in the natural association of ideas in the mind of the reader that which is expressed is so set over by way of strong contrast to that which is omitted that the contrast enforces the affirmative inference’ ” (quoting State ex rel. Curtis v. De Corps, 134 Ohio St. 295, 299, 16 N. E. 2d 459, 462 (1938)); United States v. Vonn, supra.

Strike two in this case is the failure to identify any such established series, including both threats to others and threats to self, from which Congress appears to have made a deliberate choice to omit the latter item as a signal of the affirmative defense’s scope. The closest Echazabal comes is the EEOC’s rule interpreting the Rehabilitation Act of 1973, 87 Stat. 357, as amended, 29 U.S.C. § 701 et seq., a precursor of the ADA. That statute excepts from the definition of a protected “qualified individual with a handicap” anyone who would pose a “direct threat to the health or safety of other individuals,” but, like the later ADA, the Rehabilitation Act says nothing about threats to self that particular employment might pose. 42 U.S.C. § 12113(b). The EEOC nonetheless extended the exception to cover threat-to-self employment, 29 CFR § 1613.702(f) (1990), and Echazabal argues that Congress’s adoption only of the threat-to-others exception in the ADA must have been a deliberate omission of the Rehabilitation Act regulation’s tandem term of threat-to-self, with intent to exclude it.

But two reasons stand in the way of treating the omission as an unequivocal implication of congressional intent. The first is that the EEOC was not the only agency interpreting the Rehabilitation Act, with the consequence that its regulation did not establish a clear, standard pairing of threats to self and others. While the EEOC did amplify upon the text of the Rehabilitation Act exclusion by recognizing threats to self along with threats to others, three other agencies adopting regulations under the Rehabilitation Act did not. See 28 CFR § 42.540(l)(1) (1990) (Department of Justice), 29 CFR § 32.3 (1990) (Department of Labor), and 45 CFR § 84.3(k)(1) (1990) (Department of Health and Human Services).4 It would be a stretch, then, to say that there was a standard usage, with its source in agency practice or elsewhere, that connected threats to others so closely to threats to self that leaving out one was like ignoring a twin.

Even if we put aside this variety of administrative experience, however, and look no further than the EEOC’s Rehabilitation Act regulation pairing self and others, the congressional choice to speak only of threats to others would still be equivocal. Consider what the ADA reference to threats to others might have meant on somewhat different facts. If the Rehabilitation Act had spoken only of “threats to health” and the EEOC regulation had read that to mean threats to self or others, a congressional choice to be more specific in the ADA by listing threats to others but not threats to self would have carried a message. The most probable reading would have been that Congress understood what a failure to specify could lead to and had made a choice to limit the possibilities. The statutory basis for any agency rulemaking under the ADA would have been different from its basis under the Rehabilitation Act and would have indicated a difference in the agency’s rulemaking discretion. But these are not the circumstances here. Instead of making the ADA different from the Rehabilitation Act on the point at issue, Congress used identical language, knowing full well what the EEOC had made of that language under the earlier statute. Did Congress mean to imply that the agency had been wrong in reading the earlier language to allow it to recognize threats to self, or did Congress just assume that the agency was free to do under the ADA what it had already done under the earlier Act’s identical language? There is no way to tell. Omitting the EEOC’s reference to self-harm while using the very language that the EEOC had read as consistent with recognizing self-harm is equivocal at best. No negative inference is possible.

There is even a third strike against applying the expression-exclusion rule here. It is simply that there is no apparent stopping point to the argument that by specifying a threat-to-others defense Congress intended a negative implication about those whose safety could be considered. When Congress specified threats to others in the workplace, for example, could it possibly have meant that an employer could not defend a refusal to hire when a worker’s disability would threaten others outside the workplace? If Typhoid Mary had come under the ADA, would a meat packer have been defenseless if Mary had sued after being turned away? See 42 U.S.C. § 12113(d). Expressio unius just fails to work here.


Since Congress has not spoken exhaustively on threats to a worker’s own health, the agency regulation can claim adherence under the rule in Chevron, 467 U.S., at 843, so long as it makes sense of the statutory defense for qualification standards that are “job-related and consistent with business necessity.” 42 U.S.C. § 12113(a). Chevron’s reasons for calling the regulation reasonable are unsurprising: moral concerns aside, it wishes to avoid time lost to sickness, excessive turnover from medical retirement or death, litigation under state tort law, and the risk of violating the national Occupational Safety and Health Act of 1970, 84 Stat. 1590, as amended, 29 U.S.C. § 651 et seq. Although Echazabal claims that none of these reasons is legitimate, focusing on the concern with OSHA will be enough to show that the regulation is entitled to survive.

Echazabal points out that there is no known instance of OSHA enforcement, or even threatened enforcement, against an employer who relied on the ADA to hire a worker willing to accept a risk to himself from his disability on the job. In Echazabal’s mind, this shows that invoking OSHA policy and possible OSHA liability is just a red herring to excuse covert discrimination. But there is another side to this. The text of OSHA itself says its point is “to assure so far as possible every working man and woman in the Nation safe and healthful working conditions,” §651(b), and Congress specifically obligated an employer to “furnish to each of his employees employment and a place of employment which are free from recognized hazards that are causing or are likely to cause death or serious physical harm to his employees,” §654(a)(1). Although there may be an open question whether an employer would actually be liable under OSHA for hiring an individual who knowingly consented to the particular dangers the job would pose to him, see Brief for United States et al. as Amici Curiae 19, n. 7, there is no denying that the employer would be asking for trouble: his decision to hire would put Congress’s policy in the ADA, a disabled individual’s right to operate on equal terms within the workplace, at loggerheads with the competing policy of OSHA, to ensure the safety of “each” and “every” worker. Courts would, of course, resolve the tension if there were no agency action, but the EEOC’s resolution exemplifies the substantive choices that agencies are expected to make when Congress leaves the intersection of competing objectives both imprecisely marked but subject to the administrative leeway found in 42 U.S.C. § 12113(a).

Nor can the EEOC’s resolution be fairly called unreasonable as allowing the kind of workplace paternalism the ADA was meant to outlaw. It is true that Congress had paternalism in its sights when it passed the ADA, see §12101(a)(5) (recognizing “overprotective rules and policies” as a form of discrimination). But the EEOC has taken this to mean that Congress was not aiming at an employer’s refusal to place disabled workers at a specifically demonstrated risk, but was trying to get at refusals to give an even break to classes of disabled people, while claiming to act for their own good in reliance on untested and pretextual stereotypes.5 Its regulation disallows just this sort of sham protection, through demands for a particularized enquiry into the harms the employee would probably face. The direct threat defense must be “based on a reasonable medical judgment that relies on the most current medical knowledge and/or the best available objective evidence,” and upon an expressly “individualized assessment of the individual’s present ability to safely perform the essential functions of the job,” reached after considering, among other things, the imminence of the risk and the severity of the harm portended. 29 CFR § 1630.2(r) (2001). The EEOC was certainly acting within the reasonable zone when it saw a difference between rejecting workplace paternalism and ignoring specific and documented risks to the employee himself, even if the employee would take his chances for the sake of getting a job.6

Finally, our conclusions that some regulation is permissible and this one is reasonable are not open to Echazabal’s objection that they reduce the direct threat provision to “surplusage,” see Babbitt v. Sweet Home Chapter, Communities for Great Ore., 515 U.S. 687, 698 (1995). The mere fact that a threat-to-self defense reasonably falls within the general “job related” and “business necessity” standard does not mean that Congress accomplished nothing with its explicit provision for a defense based on threats to others. The provision made a conclusion clear that might otherwise have been fought over in litigation or administrative rulemaking. It did not lack a job to do merely because the EEOC might have adopted the same rule later in applying the general defense provisions, nor was its job any less responsible simply because the agency was left with the option to go a step further. A provision can be useful even without congressional attention being indispensable.

Accordingly, we reverse the judgment of the Court of Appeals and remand the case for proceedings consistent with this opinion.

It is so ordered.


1. We do not consider the further issue passed upon by the Ninth Circuit, which held that the respondent is a “ ‘qualified individual’ ” who “can perform the essential functions of the employment position,” 42 U.S.C. § 12111(8) (1994 ed.). 226 F.3d 1063, 1072 (2000). That issue will only resurface if the Circuit concludes that the decision of respondent’s employer to exclude him was not based on the sort of individualized medical enquiry required by the regulation, an issue on which
the District Court granted summary judgment for petitioner and which we leave to the Ninth Circuit for initial appellate consideration if warranted.

2. Chevron did not dispute for purposes of its summary-judgment motion that Echazabal is “disabled” under the ADA, and Echazabal did not argue that Chevron could have made a “ ‘reasonable accommodation.’ ” App. 184, n. 6.

3. In saying that the expansive textual phrases point in the direction of agency leeway we do not mean that the defense provisions place no limit on agency rulemaking. Without deciding whether all safety-related qualification standards must satisfy the ADA’s direct-threat standard, see Albertson’s, Inc. v. Kirkingburg, 527 U.S. 555, 569—570, n. 15 (1999), we assume that some such regulations are implicitly precluded by the Act’s specification of a direct-threat defense, such as those allowing “indirect” threats of “insignificant” harm. This is so because the definitional and defense provisions describing the defense in terms of “direct” threats of “significant” harm, 42 U.S.C. § 12113(b), 12111(3), are obviously intended to forbid qualifications that screen out by reference to general categories pretextually applied. See infra, at 11—12, and n. 5. Recognizing the “indirect” and “insignificant” would simply reopen the door to pretext by way of defense.

4. In fact, we have said that the regulations issued by the Department of Health and Human Services, which had previously been the regulations of the Department of Health, Education, and Welfare, are of “particular significance” in interpreting the Rehabilitation Act because “HEW was the agency responsible for coordinating the implementation and enforcement of §504 of the Rehabilitation Act, 29 U.S.C. § 794” prohibiting discrimination against individuals with disabilities by recipients of federal funds. Toyota Motor Mfg., Ky., Inc. v. Williams, 534 U.S. 184, 195 (2002). Unfortunately for Echazabal’s argument, the congruence of the ADA with the HEW regulations does not produce an unequivocal statement of congressional intent.

5. Echazabal’s contention that the Act’s legislative history is to the contrary is unpersuasive. Although some of the comments within the legislative history decry paternalism in general terms, see, e.g., H. R. Rep. No. 101—485, pt. 2, p. 72 (1990) (“It is critical that paternalistic concerns for the disabled person’s own safety not be used to disqualify an otherwise qualified applicant”); ADA Conf. Rep., 136 Cong. Rec. 17377 (1990) (statement of Sen. Kennedy) (“[A]n employer could not use as an excuse for not hiring a person with HIV disease the claim that the employer was simply ‘protecting the individual’ from opportunistic diseases to which the individual might be exposed”), those comments that elaborate actually express the more pointed concern that such justifications are usually pretextual, rooted in generalities and misperceptions about disabilities. See, e.g., H. R. Rep. No. 101—485, at 74 (“Generalized fear about risks from the employment environment, such as exacerbation of the disability caused by stress, cannot be used by an employer to disqualify a person with a disability”); S. Rep. No. 101—116, p. 28 (1989) (“It would also be a violation to deny employment to an applicant based on generalized fears about the safety of the applicant … . By definition, such fears are based on averages and group-based predictions. This legislation requires individualized assessments”). Similarly, Echazabal points to several of our decisions expressing concern under Title VII, which like the ADA allows employers to defend otherwise discriminatory practices that are “consistent with business necessity,” 42 U.S.C. § 2000e—2(k), with employers adopting rules that exclude women from jobs that are seen as too risky. See, e.g., Dothard v. Rawlinson, 433 U.S. 321, 335 (1977); Automobile Workers v. Johnson Controls, Inc., 499 U.S. 187, 202 (1991). Those cases, however, are beside the point, as they, like Title VII generally, were concerned with paternalistic judgments based on the broad category of gender, while the EEOC has required that judgments based on the direct threat provision be made on the basis of individualized risk assessments.

6. Respect for this distinction does not entail the requirement, as Echazabal claims, that qualification standards be “neutral,” stating what the job requires, as distinct from a worker’s disqualifying characteristics. Brief for Respondent 26. It is just as much business necessity for skyscraper contractors to have steelworkers without vertigo as to have well-balanced ones. See 226 F.3d, at 1074 (Trott, J., dissenting). Reasonableness does not turn on formalism. We have no occasion, however, to try to describe how acutely an employee must exhibit a disqualifying condition before an employer may exclude him from the class of the generally qualified. See Brief for Respondent 31. This is a job for the trial courts in the first instance.