with whom Mr. Justice Rehnquist joins, concurring in part and dissenting in part.
I join Part IV of the Court’s opinion; as to Parts I, II, and III, I dissent.
Gender-based actuarial tables have been in use since at least *7261843,1 and their statistical validity has been repeatedly verified.2 The vast life insurance, annuity, and pension plan industry is based on these tables. As the Court recognizes, ante, at 707, it is a fact that “women, as a class, do live longer than men.” It is equally true that employers cannot know in advance when individual members of the classes will die. Ante, at 708. Yet, if they are to operate economically workable group pension programs, it is only rational to permit them to rely on statistically sound and proved disparities in longevity between men and women. Indeed, it seems to me irrational to assume Congress intended to outlaw use of the fact that, for whatever reasons or combination of reasons, women as a class outlive men.
The Court’s conclusion that the language of the civil rights statute is clear, admitting of no advertence to the legislative history, such as there was, is not soundly based. An effect upon pension plans so revolutionary and discriminatory — this time favorable to women at the expense of men — should not be read into the statute without either a clear statement of that intent in the statute, or some reliable indication in the legislative history that this was Congress’ purpose. The Court’s casual dismissal of Senator Humphrey’s apparent assumption that the “Act would have little, if any, impact on existing pension plans,” ante, at 714, is to dismiss a significant manifestation of what impact on industrial benefit plans was contemplated. It is reasonably clear there was no intention to abrogate an employer’s right, in this narrow and limited context, to treat women differently from men in the face of historical reliance on mortality experience statistics. Cf. ante, at 713 n. 25.
The reality of differences in human mortality is what mortality experience tables reflect. The difference is the added *727longevity of women. All the reasons why women statistically outlive men are not clear. But categorizing people on the basis of sex, the one acknowledged immutable difference between men and women, is to take into account all of the unknown reasons, whether biologically or culturally based, or both, which give women a significantly greater life expectancy than men. It is therefore true as the Court says, “that any individual’s life expectancy is based on a number of factors, of which sex is only one.” . Ante, at 712. But it is not true that by seizing upon the only constant, “measurable” factor, no others were taken into account. All other factors, whether known but variable — or unknown — are the elements which automatically account for the actuarial disparity. And all are accounted for when the constant factor is used as a basis for determining the costs and benefits of a group pension plan.
Here, of course, petitioners are discriminating in take-home pay between men and women. Cf. General Electric Co. v. Gilbert, 429 U. S. 125 (1976); Nashville Gas Co. v. Satty, 434 U. S. 136 (1977). The practice of petitioners, however, falls squarely under the exemption provided by the Equal Pay Act of 1963, 29 U. S. C. § 206 (d), incorporated into Title VII by the so-called Bennett Amendment, 78 Stat. 257, now 42 U. S. C. § 2000e-2 (h). That exemption tells us that an employer may not discriminate between employees on the basis of sex by paying one sex lesser compensation than the other “except where such payment is made pursuant to ... a differential based on any other factor other than sex . . . .” The “other factor other than sex” is longevity; sex is the umbrella-constant under which all of the elements leading to differences in longevity are grouped and assimilated, and the only objective feature upon which an employer- — or anyone else, including insurance companies — may reliably base a cost differential for the “risk” being insured.
This is in no sense a failure to treat women as “individuals” in violation of the statute, as the Court holds. It is to treat *728them as individually as it is possible to do in the face of the unknowable length of each individual life. Individually, every woman has the same statistical possibility of outliving men. This is the essence of basing decisions on reliable statistics when individual determinations are infeasible or, as here, impossible.
Of course, women cannot be disqualified from, for example, heavy labor just because the generality of women are thought not as strong as men — a proposition which perhaps may sometime be statistically demonstrable, but will remain individually refutable. When, however, it is impossible to tailor a program such as a pension plan to the individual, nothing should prevent application of reliable statistical facts to 'the individual, for whom the facts cannot be disproved until long after planning, funding, and operating the program have been undertaken.
I find it anomalous, if not contradictory, that the Court’s opinion tells us, in effect, ante, at 717-718, and n. 33, that the holding is not really a barrier to responding to the complaints of men employees, as a group. The Court states that employers may give their employees precisely the same dollar amount and require them to secure their own annuities directly from an insurer, who, of course, is under no compulsion to ignore 135 years of accumulated, recorded longevity experience.3
See H. Moir, Sources and Characteristics of the Principal Mortality Tables 10, 14 (1919).
See, e. g., United Nations, 1970 Demographic Yearbook 710-729 (1971).
This case, of course, has nothing to do with discrimination because of race, color, religion, or national origin, cf. ante, at 709, and nn. 15 and 16. The qualification the Bennett Amendment permitted by its incorporation of the Equal Pay Act pertained only to claims of discrimination because of sex.