(FORTUNE Magazine) – Baby-boomers have ushered in most every major trend over the past 50 years. But it was their grandparents who initiated the most radical demographic change of the past half-century--a dramatic decline in death rates at older ages. In fact, about the time boomers were rambunctiously burning draft cards, their elders quietly began nullifying actuarial tables. By 1990 there were more than 1.5 million Americans age 85 and over who wouldn't have been alive if death rates had stayed at the 1960 level.

Extrapolating this trend, demographer James Vaupel has made a bold prediction: Half of the girls and a third of the boys recently born in the developed world will live to be 100. Vaupel similarly expects millions of former flower children to defy federal population forecasts and make good on their old chant, "Hell no, we won't go!"--he has projected there could be nearly 37 million boomers age 85 and over by 2050, more than twice the government's best guess. That would mean a much higher proportion of senior citizens nationwide than Florida has today. It would also mean that we should now be considering far more stringent policies to prevent future economic crises than are on the table, from raising the retirement age to trimming Social Security and Medicare entitlements.

Vaupel, who heads Germany's Max Planck Institute for Demographic Research, is more optimistic than most demographers about the prospects for longer life. But he's no shallow visionary. A few years ago many of his colleagues scoffed when he challenged a grim canon about aging. It holds that death rates rise exponentially with age in adult animals, including humans--the older you are, the theory goes, the more likely you are to die. Aided by other researchers, he marshaled data on everything from Swedish women to Medflies to show it ain't so; for good measure, he threw in supporting data on the death rates of old cars. The team demonstrated that mortality can plateau and, strangely, even drop among the very old--as if the Fates were nodding off after a long wait.

Vaupel sees this "mortality deceleration" as a subplot of a grand mystery that has preoccupied demographers for over a decade: Why have the elderly been living longer than their forebears since about 1970? Some of the causes are obvious, such as the averting of millions of fatal heart attacks by blood-pressure drugs widely used since the 1960s. But many experts on aging feel that such well-known factors can't explain the trend's surprising speed and breadth. For instance, the number of centenarians in various developed countries has doubled every decade since the 1960s.

Casting about for explanations, some demographers theorize that deep, little-understood changes are afoot that will help sustain the trend for decades. Vaupel has stuck his neck out farther than most by proposing that the aging process may actually slow down in very old people, an idea based on his mortality-deceleration work. That particular idea remains highly controversial. But Vaupel's bullish view that longevity gains will continue apace is widely shared. Indeed, many demographers are now more bullish than the Social Security Administration, which projects that the decline in old-age death rates will slow to a crawl early in the next century.

The bulls' predictions raise a burning issue: If we receive a gift of extra years, will it turn out to be a Pandora's box filled with hobbling diseases? For most of this century death rates and the prevalence of chronic diseases among the elderly have dropped in tandem. But "we're balanced on a razor's edge," says Eric Stallard, a demography professor at Duke University. If medical advances make mortality fall faster than disease, we'll wind up spending costly extra years in nursing homes. Or worse: "We may face the gruesome prospect of poor, disabled, homeless older Americans living out the end of their lives on city streets and in parks," warns Edward L. Schneider, dean of gerontology at the University of South California.

Vaupel, a U.S. citizen reared on Long Island, was one of the first researchers to issue wake-up calls about the new demography of aging, as he calls it. The implications came home to him when his daughter, Anna, was born in 1984. Musing about her future, he calculated that if the death-rate declines observed since the 1960s continued, girls in her "birth cohort" could expect, on average, to become centenarians. Since then, he's moderated his forecast some. "My best guess now is that 50% of newborn white females in the developed world will celebrate their 100th birthdays," he says.

Vaupel's projections are still far more bullish than the Social Security Administration's midrange forecasts, which are typically used to project Uncle Sam's liabilities. According to the agency, U.S. life expectancy, or average life span at birth, will rise to about 82 years in 2075, from about 76 now. That's roughly a fourth as much as it rose over the past 75 years.

As Vaupel has projected, a heftier rise in life expectancy would increase by tens of millions the number of boomers who keep on trucking after 2030. That has huge fiscal implications. Federal medical payouts for individuals over 84 are more than twice as high, on average, as for those between 65 and 69. If numbers aren't your cup of tea, just picture the budget circa 2030 as a snake trying to swallow a cow.

Like all long-term projections, forecasts of mortality patterns over the next century are no more than educated guesses. Underscoring this uncertainty, some demographers flatly disagree with Vaupel and predict the bull market for longevity gains is about to run out of steam. The most vociferous of these bearish experts is the University of Chicago's Jay Olshansky. An affable man known for ripping the bulls in technical journals, Olshansky cheerfully argues that we'll all go under in a timely way from the federal government's point of view--he feels the Social Security Administration's intermediate projections are about right. "The practical limit for life expectancy hasn't changed at all. It's still about 85," he declares.

Pushing life expectancy into the upper 90s, which some bulls think is possible over the next century, would be like human sprinters getting up to cheetah speed--the human body isn't physically capable of doing that in the foreseeable future, Olshansky contends. When sparring with Vaupel in the Boston Globe a few years ago, he projected that his daughter, Jessica--who happens to be the same age as Vaupel's daughter--was born into a generation with a life expectancy of 85.

At the bottom of the bull-bear rift are radically different views on the future and the past. The bulls are more sanguine about medical science's ability to keep the Reaper in retreat in coming decades. They envision major life expectancy gains from advances such as artificial organs and medicines that cure major killers, and they're willing to entertain hopes of drugs that will someday slow the aging process--an idea that's gaining credibility, thanks to the discovery of gene mutations that can dramatically extend lower animals' life spans (see part one of this series, "Live a Lot Longer," in the fortune.com archive). The bulls also have a rosier view of the forces that have raised U.S. life expectancy by a whopping 30 years since 1900. One bullish theory posits that good health early in life has cut the risk of various chronic diseases of aging. That implies gains against such diseases should continue well into the next century, regardless of what high-tech medicine does for us.

The developed world's life expectancy pattern over the past century falls roughly into three parts. Before 1950 it rose quickly, thanks to the advent of antibiotics, better nutrition, and other factors that mainly reduced mortality among the young--in particular, infant mortality. After much of that low-hanging fruit was plucked, life expectancy gains slowed and seemed to plateau in the 1950s and '60s (see charts). Then around 1980 it became apparent that life expectancy was rising again. This time, though, stellar gains were occurring among older people.

Before that pattern rose like an uncanny phoenix from worldwide death statistics, many experts believed we were approaching an intractable biological limit on life expectancy, says John Wilmoth, a demographer at the University of California at Berkeley. In 1978, for instance, a prominent French demographer argued that life expectancy couldn't exceed 73.8 years for men and 80.3 years for women. Four years later life expectancy in Japan barreled through the presumed male limit, and in 1985, through the female one.

The idea of an 85-year limit remains viable. Japanese women, the current world champs for longevity, have a life expectancy of about 83. Further, "the rise in life expectancy is decelerating," says Olshansky.

That's true. The pre-1950 gains were faster than the gains after 1970, thanks to factors that cut infant mortality. But the bulls say Olshansky is focusing on the wrong indicator. What's important, they argue, are falling death rates at older ages, a profound change that's largely masked when factored into life expectancy averages--just as broad stock-market indexes barely registered skyrocketing Internet stock prices, which some observers feel is a harbinger of fundamental economic change.

In a 1994 study, Vaupel and colleagues showed that average annual mortality declines at age 80 and above actually sped up between 1970 and 1990 in nine developed nations. Such data, says Berkeley's Wilmoth, suggest that "if there's a biological limit to human longevity, we're not getting close to it yet."

As Vaupel racked up debating points, French centenarian Jeanne Calment emerged as the bulls' Michael Jordan. "When she got to 115, the naysayers [about continuing death-rate declines among the aged] said she probably wouldn't live another year," says Stallard, the Duke professor. The awesome little old lady blew right on by--she died in 1997 at age 122 and five months.

"I have been strongly impressed by the life of Madame Calment," wrote Berkeley demographer Kenneth Wachter shortly before her death. Among other things, he said, Calment had helped convince him that mortality really does decline among the very old--just as Vaupel had said.

But while Vaupel has convinced colleagues that mortality deceleration is real, its cause remains a subject of heated debate. Bears like Olshansky argue that the phenomenon is a trivial statistical artifact--death rates, which tend to rise as the weaker members of a population die off, slow when only a robust few remain. Indeed, he notes, "the death rate for people over 120 was zero shortly before Jeanne Calment died"--she was the only one in that age category. Vaupel acknowledges that this "heterogeneity of frailty" among aging individuals might explain mortality deceleration. On the other hand, he says, it might also involve "behavioral and physiological changes with age."

Consider automobiles, he says. Using car-registration data, which show when autos are no longer in use and presumed dead, Vaupel demonstrated that death rates among Toyotas and Chevrolets tend to plateau after about a decade of use--their mortality curves resemble the ones for Medflies and people. This is partly due to heterogeneity of frailty--lemons die fast, leaving nonlemons to age more gracefully. But it's also likely, says Vaupel, that "older cars are repaired, and driven less frequently and more carefully. As cars age, their owners may also age and become more stolid and responsible." By analogy, "it's possible some of the mortality deceleration we've seen is due to some slowing of the aging process at higher ages"--an effect perhaps tied to good genes and a knack for preserving the old bod.

Unimpressed by all this, Social Security officials for years have ignored outsiders' recommendations, including Vaupel's, to issue more bullish mortality projections. Because of the agency's conservative mortality forecast, a 2.2% payroll tax increase that it estimates is needed to keep its trust fund actuarially sound through 2073 is more than 30% below what it should be, estimates Ronald D. Lee, an influential Berkeley demographer regarded as a moderate. Some demographers say privately that the agency is under pressure to maintain its cautious stance, lest its projections force policymakers to consider politically untenable tax hikes.

Not true--politics doesn't affect the Social Security Administration's forecasts, asserts Harry Ballantyne, the agency's chief actuary. Its intermediate forecast, he adds, recently was deemed "well within the range considered reasonable" by 41 experts convened by the Society of Actuaries, a professional group.

A flap in Canada, however, showed how hard it is for government actuaries to insulate themselves from politics. Last August, Bernard Dussault, Canada's chief actuary, was fired while preparing a report on the fiscal health of his government's retirement fund. Dussault charged that he was dismissed after refusing an order to alter a figure in the report that would have embarrassed the government--an assertion disputed by his former boss, who told reporters that Dussault was fired because of "management differences." In December, Dussault's replacement concluded the pension plan remained strong and workers wouldn't have to pay more--just as they had been promised earlier by the Finance Minister.

Projecting Medicare costs raises even thornier issues than do Social Security forecasts, including the one we care about most: If we live longer, will our extra years be spent in good health?

A 1997 study at Duke offered good news: Disability rates among Americans over 64 dropped, on average, 1.3% a year between 1982 and 1994. Even better, the gains against the Reaper's vanguard of imps were growing--after declining 1.1% a year until 1989, disability fell an astonishing 1.5% annually over the next five years.

Predictably, the bulls and bears disagree about this. No one doubts that medical advances, such as cardiac-artery bypass operations to ward off disabling heart attacks, played a role. Healthier lifestyles clearly did too--men who exercise, keep their weight down, and don't smoke apparently postpone the onset of "functional disability" by about five years, compared with couch potatoes who smoke, according to a study last year by Stanford University researchers.

But there's sharp disagreement about whether the gains are the beginning of a long-term trend. Pushing life expectancy past 85 with medical advances gives us "manufactured time," says Olshansky. It may merely shift "the burden of senescence" to costlier chronic diseases--a person may avoid a fatal heart attack at 75, say, only to die at 85 after an eight-year struggle with Alzheimer's disease. He adds that new diseases of aging may appear as we smash into "hazards lurking in the older regions of the human life span." Such hazards may include harmful genes that come into play only late in life, like the one that causes Huntington's disease, a fatal brain disorder whose symptoms don't appear until midlife.

At this point in the debate, the bulls usually trot out some of their most heartening and surprising data: Chronic-disease risks among the elderly appear to be dropping in part because of improved nutrition and fewer infections early in the century. British researcher David Barker, for instance, has published studies indicating that low birth weight, a sign of possible fetal malnutrition, increases the risk of heart disease and other late-life disorders. An analysis of records kept by the Union Army pension program, set up in 1862, showed that early-life infections among Civil War veterans predisposed them to higher risk of heart-valve disease, clogged arteries, and other diseases of aging.

The bulls see the latter study, reported in December by Massachusetts Institute of Technology economist Dora Costa, as a slam dunk. "Chronic diseases shouldn't necessarily be thought of as part of the natural process of aging," Costa concluded. If she's right, baby-boomers may well age more gracefully than earlier generations, since antibiotics let them grow up with fewer serious infections.

The bulls note that boomers have at least two other things going for them: Over a fourth of them attained a bachelor's degree or higher, which is strongly correlated with healthier lifestyles--a recent study on smokers who had heart attacks, for instance, showed that nearly 90% of college grads quit afterward, while only 10% of those with less than a high school education did. And because boomers are expected to live longer, medical costs during the last year of life, when a person's Medicare expenses typically soar, may actually drop. A 1993 study showed that average final-year spending totals $15,500 for people who die between 65 and 74, ages at which doctors are willing to try heroic measures, compared with less than $9,000 for those who die after 89.

The final score in our momentous game of life and delayed death is still very much up in the air. Vaupel, for one, is optimistic about the outcome--he believes better health at older ages is largely driving the mortality declines after 65, so our average life spans and "health spans" will continue to rise in tandem. For his part, Olshansky says that to increase health spans, "we urgently need to put more resources into attacking diseases that cause frailty and disability." On that point, at least, the bulls quite agree.