WOMEN, ON THE AVERAGE, live longer than men. This means, among other things, that women pay lower rates for life insurance and receive lower pension payments upon retirement. Those financial consequences of women's longer life expectancy have been accepted as legitimate because they are based on sound statistics. But if you scrutinize the implications of those statistics, as the Supreme Court did the other day, you begin to wonder whether it is fair to make so much, in public policy and economic planning, of this one generalization about life expectancies when other generalizations that could be made (by race, ethnic background, family medical history or personal habits) are not applied.

This question has been raised, almost casually, as part of the discussion over the Equal Rights Amendment and women's rights in general. But is has now been opened up for public debate by the court's decision that the Civil Rights Act of 1964 bars employers from requiring female workers to contribute more money than male workers to the same pension program. The immediate impact of the decision is narrow - only a handful of employers, most of them units of government, have pension programs of that kind. The implications are wide, however, because the same rationale could knock out pension plans that pay retired women less than retired men. That rationale also raises questions, at least at the level of public policy, about other economic arrangements in which gender is used as the dividing line between groups of people.

Justice John Paul Stevens, writing for the court's majority, noted that life-expectancy tables have also been worked out on the basis of race and national origin. Yet these tables are not used to justify higher insurance premiums and higher pension benefits for blacks than for whites in the same way that the tables based on gender are used to justify higher premiums and payments for men than for women. Indeed, some states outlawed higher insurance rates for blacks in the 19th century. Similarly, life expectancy tables have been developed for smokers and nonsmokers and probably could be developed to distinguish between various occupational groups or between those with some family history of cancer and those without.

The point is not that classification by group for insurance, pension and other programs is inherently bad. Some classification is necessary to make group plans work at all. Those based on race are generally regarded as unfair and unacceptable, both in public policy and constitutional terms. Those based on gender are not yet so regarded. By reading broadly on the anti-sex discrimination provision of the Civil Rights Act, the court has taken a step toward ending this inconsistency. And by challenging the coventional wisdom that women's longer life expectancy automatically justifies certain economic consequences, the court has raised serious questions about the premises on which almost all of the nation's insurance programs and many pension systems rest.