Projected life expectancy has taken a jump of more than three years--the biggest leap since that statistic was first recorded--which has set statisticians and actuaries to revising their economic and social projections, according to a panel meeting at the American Association for the Advancement of Science here.

Until recently, the Census Bureau and other government agencies had stuck to their longtime prediction of human life expectancy for up to the year 2000 as being about 70 for men and 78 for women.

But Eileen Crimmins, a gerontologist with the University of Southern California, and others now project a sudden increase in the human life span and have criticized the government for not taking the new projections into account. Crimmins estimated that life expectancy in the year 2000 will be about 74.3 years for men and 86.2 years for women. Other statisticians have slightly lower estimates.

The numbers may continue to rise into the 21st Century, eventually reaching a supposed maximum human life span of about 95 years, the panel said.

These predictions alter beliefs of the past two decades, during which statisticians regarded age 70 to 75 as the likely limit on human life spans for years to come.

But over the past decade, an unexpected decline in many major diseases--including heart disease, stroke, diabetes and pneumonia--among those over 65 has forced researchers to revise estimates of population and longevity.

"This is the biggest change ever recorded," Crimmins said. The government statisticians "thought that it couldn't happen. They thought that 70 was about as long as we were going to live." They expected that chronic diseases such as heart disease and stroke would not yield soon to treatment, she said.

At a session today on aging and mortality, Crimmins said that the government is underestimating the number of people over 65 who will be alive in the year 2000 by about 15 percent, or more than 4 million people.

Recent government studies have begun to adjust the numbers upward, but still not to the levels given by other statisticians, she said.

Barbara Boyle Torrey and Douglas Norwood, Office of Management and Budget economists, presented data showing that, even under the old figures, which are thought to undercount those over 65, the disproportionate ratio of old-to-young in the country would mean a loss to the federal treasury of tens of billions of dollars in the year 2000.

Although the OMB economists were not speaking for their agency, they said that federal agencies must begin to take into account for policy planning the projections of an older population. Norwood said that federal retirement plans for the most part do not pay attention to the projections of increased longevity and an older population.

Torrey said that among the most interesting data in her study of how the old-age boom will affect the federal treasury was the finding that older people do not drain the treasury the most, because they have spent years contributing to it. However, on average they will draw more than they put in, because of increasing life spans.

Torrey said that those who die young are the greatest loss to the treasury because of lost tax revenue.

Leonard Hayflick, a gerontologist at the University of Florida, said the graph of human life expectancy is beginning to look more and more "square."

That is, "about 100 percent of the people are born and live, and then on the stroke of midnight on their 95th birthday they all drop dead. That is not as facetious as it sounds, that is in fact" the way the life expectancy curve is beginning to look, he said.

That scenario includes the elimination of disease for the most part, and death coming because of natural biological breakdown of the system at about age 95.