(FORTUNE Magazine) – ENTER the strange world of Social Security, as exotically inside out as the domain of black holes and anti-matter that physicists describe. Within the borders of this unexpected land the U.S. runs not one of its familiar handwringing deficits but surpluses, accumulating at the pace of $109 million a day. Over the past few months surplus fever has broken out. Politicians, economists, and editorial writers have marveled at the size of the Social Security reserve and gleefully proposed plans to spend it. Some want to use it to solve America's problems: to vaporize the budget deficit, educate the young, build housing for the homeless, or satisfy all the needs of the elderly. Others want to leave it alone, thus prepaying the retirement of the baby boom. As prodigious as the reserve is, it cannot meet all those needs. The struggle for control over that vast sum -- a total of nearly $12 trillion at its peak in the year 2030, according to current projections -- will be one of the most important political and fiscal issues from now until the turn of the century. The fundamental issue is as much ethical as economic: How much does the ! country want to do for its elderly and how rich, or how poor, does it want to leave its children? In this article FORTUNE argues that under the present system, the Social Security reserve will grow too big. Part of it should be siphoned off to cut down the size of the federal budget deficit. We also propose changes in the system that would trim benefits, but for future, not present, retirees. The proposition here is for a reserve that will increase savings and thus strengthen the economy, and that will also safeguard the retirement of baby-boomers as well as the generation that comes after them. Unless that is taken care of, the reserve, vast as it is becoming, will disappear by 2050, leaving the politicians scrambling to avoid bankruptcy, as they did when they last fixed the Social Security system in 1983. What's creating the surpluses is plain enough. To keep Social Security solvent, Congress nudged up the Federal Insurance Contributions Act (FICA) tax -- colloquially the payroll tax -- to 15.2% on the first $45,000 in income, half paid by workers and half by employers. Because the baby-boomers, 75 million Americans born between 1946 and 1964, have reached working age, the tax is pumping money into the Treasury at a wondrous rate, about $317 billion this year. Even though some people labor under the illusion that the money taken out of their pay is set aside for them, Social Security is a pay-as-you-go system in which money is transferred from working stiffs to retired folks. The Treasury will send $277 billion or so of this year's income to current beneficiaries. BY LAW the surplus $40 billion must be invested in special government bonds that pay the same interest as other long-term federal debt but cannot be traded and can always be redeemed at par value. At the end of this year the Social Security trust funds will have a reserve -- the accumulation of annual surpluses -- of $109 billion of such IOUs, collecting interest at 10% to 11%. The Treasury lends the cash that these IOUs represent to the Pentagon, the Department of Agriculture, and other federal agencies to help pay their bills. Because of the low birth rates of the 1930s and early 1940s, the number of retirees and their dependents will remain small for the next few decades. So for a while the funds will grow robustly. Figuring how big they will get is largely a matter of guessing such variables as fertility, mortality, immigration, and productivity rates. So Social Security actuaries prepare a high forecast (Alternative I), a low one (Alternative III), and a pair of Alternative IIs in between. It is Alternative IIB, the one the Social Security trustees consider most likely, that predicts the old age and disability surpluses will have heaped to a mammoth reserve of $12 trillion by the year 2030. Taking account of inflation (four decades at 4% a year), that translates into $2.3 trillion of 1988 dollars, nearly as big as the current national debt. Stack that $12 trillion up against another important measure as well. Most students of the Social Security system agree that the funds ought to have a contingency reserve, a cushion in case of recession, about the size of a year's payouts. Even this year's $109 billion reserve doesn't match the standard. But $12 trillion equals 3 1/2 times the anticipated commitments of 2030. After 2020 so many baby-boomers will have retired that the income from payroll taxes will no longer be enough to support retirees. The trust funds will continue to grow for a decade because of interest income. Then the funds will have to start cashing in the IOUs they received when they lent the money to support federal spending. To pay up, the government will have to raise taxes or borrow, delaying the day of reckoning. Just as the boomers were preceded by a baby bust, they are followed by one. The burden on that second wave of busters and following generations could be crushing, especially if the future is darker than Alternative IIB. A. Haeworth Robertson, chief actuary for the Social Security Administration in the late 1970s, thinks his colleagues are too optimistic in assuming that women will have as many as 1.9 children in the future. ''I consciously set out to have more than two children as a matter of duty,'' says Robertson, not a whimsical man. ''People don't do that anymore.'' CURRENTLY the fertility rate is a little less than 1.9, and if the proportion of women entering the labor force increases, it could drop further. A fall in fertility, accompanied by even a small slip in productivity, would mean that workers 50 years from now might have to sacrifice 15% or even 20% of their pay (and employers a like amount) to cover Social Security's promises. It is not hard to imagine angry confrontations -- tax wars perhaps -- in which the generations do battle over whether the young should pay more or the old should take less. One solution might be to have baby-boomers volunteer to forgo Social Security and take more responsibility for their old age. Pierre du Pont, briefly a candidate for the Republican presidential nomination, suggested that government give workers a $1 tax credit for every $1 they put in a financial security account, which would buy stock and other investments. The idea is appealing in some respects, but in its early years, at least, it would cost the Treasury billions in lost taxes that it can't afford. Another way to ward off future generational warfare, suggested by Budget Director James C. Miller, would be for Congress to let the trust funds buy a broad portfolio of corporate bonds. ''With government securities you have to rely on the willingness of future taxpayers to make good,'' says Miller. ''With corporate bonds you have something substantial.'' But under Miller's proposal, the Treasury would still have to sell bonds to cover current spending, obliging taxpayers to settle up eventually. A better plan for financing the future might be to change the way the Social Security surpluses are used. Now the Treasury lends them to government agencies for consumption. A smarter alternative would be to reduce their size by using some of the payroll tax revenues to cut the federal deficit. The rest -- about $1 trillion over the years -- would be lent to the government to reduce the federal debt. Paying down the debt increases national savings. Think of that $1 trillion as a kind of national nest egg. Through the 1950s the U.S. net savings rate averaged over 7%, but in the early Eighties it slumped to 3.4% and in the past couple of years it fell to about 2%. Over that same period the payroll tax, employers' and employees' contributions combined, rose from 3% in 1950 to 15.2% currently. From an economic perspective, it doesn't make much difference who saves, government or individuals. Either way society benefits. As the supply of savings expands, interest rates tend to fall, other things being equal. Lower rates encourage businesses to borrow for factories and new machines. Productivity, the output per worker, grows. With more goods and services available as a result, Americans grow richer. No wonder economists love nest eggs. THE QUARREL is between those who want a big egg and those who prefer a smaller one. In this article, FORTUNE will make a case for a modest-size egg. Party allegiances don't necessarily determine egg preferences. Vice President George Bush consults economists Martin Feldstein of Harvard, who first / proposed the big egg in 1974, and Michael Boskin of Stanford, who thinks a large egg attractive but impractical. The Vice President has not decided how he likes his eggs. The Dukakis team appears to be more united behind a big egg, but is not yet firm on how big. Among Democrats, Senator Daniel Patrick Moynihan of New York is by far the most vocal salesman for extra large. But the bigger the egg, the more pressure on Congress and the President to crack it open to pay for new programs. ''I am deeply concerned about the ability of our political system to keep the money,'' says Boskin. Congressman Andrew Jacobs of Indiana, chairman of the House Social Security subcommittee, puts it more colorfully : ''It'll be like walking through a bad neighborhood with a diamond ring.'' Think of that wily champion of the elderly, Representative Claude Pepper of Florida, who recently maneuvered an expensive proposal onto the floor of the House without a hearing. Pepper's bill would have provided long-term home care for the disabled of all ages and would have raised the money by increasing payroll taxes a bit for those earning more than $45,000 a year. The House defeated the bill this time, but the fatter the Social Security reserve gets, the harder it will be to convince the old that the dollars should be saved for future generations. Not just the old will stake claims. Isabel Sawhill, an economist at the Urban Institute, notes that there are many ways to use surpluses to promote growth. ''We can invest in human capital by paying for education with surpluses,'' she says. Representative Bill Green of New York wants to use the money for housing projects and other public works. According to the law of expanding appetites, the more money available, the more imaginative the spending schemes. For example, in his worried examination of a declining population, The Birth Dearth, Ben J. Wattenberg suggests that government promote fertility by giving parents $2,000 a year for every child they have under 16. ''Such big money can and should come from only one source,'' he concludes, ''the Social Security trust funds.'' Three economists at the Brookings Institution, Henry Aaron, Barry Bosworth, and Gary Burtless, propose a big Social Security reserve with a difference. Their aim is to build up the $12 trillion trust fund and balance the budget at the same time. ''We have a golden opportunity to raise our national savings rate,'' says Bosworth, ''so let's put our money where our mouth is.'' / The Brookings trio is circulating among colleagues a 100-page draft of its plan. It calls for an end to the double talk by which Social Security is legally separate from the rest of the federal budget, yet is included in the total, or so-called unified budget, when it comes to meeting the deficit reduction targets of the Gramm-Rudman-Hollings law. Without aid from the Social Security surplus, the $1.1 trillion unified budget would likely come up $197 billion short this year instead of the $157 billion the Congressional Budget Office projects. Gramm-Rudman requires that the budget be balanced by 1993, and the $97 billion surplus Social Security will run that year would be a big help. Brookings insists that Congress should balance the budget, or at least come close, without resorting to the Social Security surplus. This would mean that the government would no longer be borrowing from the trust funds to cover current spending, so it could use the money to retire old debt. When U.S. Treasury bonds held by Sumitomo Life Insurance, say, come due, the Treasury would pay them off instead of rolling them over. Sumitomo, then, would likely put its money in the private sector, stimulating the economy and pumping up investment. The trust funds would still get nothing more than an IOU from the government. And the government would still have to raise taxes eventually to cash in the IOUs to support the aged baby-boomers. But here's the wonder of the Brookings plan: Workers will not object to higher taxes when the time comes, says Brookings, because the boost in capital formation would increase productivity and, as a result, their wages. ''The numbers just happen to work out perfectly,'' Bosworth maintains. ''With just a couple of adjustments in payroll taxes we can finance the country's retirements indefinitely.'' BY AND LARGE, other economists don't question the Brookings calculations. As theory the plan is elegant. But it is also somewhat arbitrary, utopian, and risky. One problem: The workers of the future who enjoy higher wages will not necessarily be any happier with higher taxes than today's highly paid employees. Fiscal drag is the immediate threat. The Brookings plan requires a swing of over $250 billion from deficit to surplus in the unified government accounts over six years, from $157 billion in the red in 1988 to $97 billion in the black in 1993. Brookings has no formula for achieving that dramatic shift. ''From an economist's point of view it doesn't make any difference whether you cut spending or raise taxes,'' maintains Bosworth, although not all his fraternity would agree. Presumably the political solution would require a mixture of both. As the government takes money out of the hands of workers and businesses for taxes, the economy would tend to weaken. In the Brookings model that loss of strength would be offset by the stimulus of lower interest rates and the influx of capital to the private sector. But the world is not tidy. Says Todd May, FORTUNE's chief economist: ''When you take money out of people's pockets the impact on the economy is fast and direct. It isn't nearly so clear that interest rates will come down enough and capital flows will increase enough to compensate.'' If the country were to slip into a recession, employment would fall and pull down payroll taxes and Social Security surpluses with it. To their credit Bosworth and the rest of the Brookings bunch acknowledge what Moynihan and other politicians are inclined to brush aside: One of the three Social Security trust funds supported by the payroll tax will soon spring a leak. Two of the funds are surplus storehouses -- the old age and survivors insurance fund and the disability insurance fund. But the third, Part A of Medicare, which covers hospital stays for the elderly and disabled for up to 60 days, is expected to go bankrupt early in the next century. Projections show that by about 2005 the fund will run a $40 billion deficit. (Part B of Medicare, which pays the doctors' bills of the elderly, is in no such danger because, incongruously, most of the money comes from general tax revenues.) Brookings figures that to plug the leak in the hospital fund and keep the reserve growing, workers and employers will have to kick in an extra couple of percentage points on their payroll tax. Under the Brookings plan, the boomers would be the first generation stuck with two Social Security bills, their parents' and part of their own. Boomers who are now working make up half the labor force and are paying for the retirement of their parents' generation. The young have paid for the old ever since the first beneficiary, Ida Fuller, a Vermont law clerk, quit work in 1940. From the time the system began collecting taxes in 1937, she had paid a total of $22, and her employer another $22. Ida lived long enough to collect more than $20,000. Someone who retires at 65 these days will recover within four years all the money he and his boss paid into the system.

BESIDES having the boomers support current beneficiaries, the Brookings plan would require them to build the national nest egg that will help finance their own golden years as well. They would have to contribute higher personal taxes to balance the budget and greater payroll taxes to make up for the leaky hospital fund. Why should they take the double hit? True, the boomers are so numerous that they could be unfairly burdensome for the following generation of baby-busters. But Rudolph Penner, former director of the Congressional Budget Office, points out what is often overlooked: The busters will almost certainly be richer than their elders. Productivity is increasing, slowly to be sure, but still increasing. As long as that is the case and business keeps sharing the gains with labor, real wages will continue to rise. If productivity continues to go up 1.6% a year, the recent average, wages could double by 2032 (after excluding inflation). The single expense that could make busters poorer than their forebears, Penner is quick to acknowledge, is too big a Social Security tax. The trick is to balance the load between the generations so that the busters carry some responsibility for the health and comfort of their elders, but end up at least as wealthy. That calls for moderation: a medium Social Security reserve that would neither choke off economic growth nor create as much temptation as the current projected surplus or as much pain as the Brookings plan. The dimensions of the ultimate egg don't have to be calibrated just yet. But a sensible midterm goal might be a $1 trillion Social Security reserve, in nominal dollars, by 2010 -- the year before the oldest baby-boomers reach 65. That assumes that the rest of the federal budget will have been in balance since the mid-Nineties, or the reserve won't have its salubrious effect of lowering interest rates. No magic attends that $1 trillion figure, and it can be moved up or down as the economy changes speed. But for several reasons it's an attractive target. In rough numbers, piling up that kind of reserve over 15 years or so would require an average Social Security surplus of about $65 billion a year, although much less in the early years and more in the later ones, when inflation and economic growth will make $65 billion look a lot smaller. Yet $1 trillion would add substantially to the stock of national savings. And as a big round number it makes for good political slogans: The country needs a $1 trillion savings account to get the baby-boomers started on retirement. Also, $1 trillion is close to a year's worth of Social Security expenses in 2010, so the system would have a respectable contingency fund. A sustained $1 trillion reserve, plus other changes in the system, will help provide for the elderly long after 2020, the first year that Social Security payouts exceed payroll taxes. To begin to reach those twin objectives -- a balanced budget and a $1 trillion nest egg -- Congress ought to do what Martin Feldstein recommends: charge the Social Security trust funds for doctors' bills for the elderly (Medicare's Part B), just as it charges the funds for hospital bills (Part A). Like hospital care, doctors' care is part of security for the old and should be paid out of the trust funds, not general revenues. THAT WOULD save the rest of the budget $26 billion this year and more in years to come. While Congress defeated Pepper's home care bill, it passed, and President Reagan signed, catastrophic illness legislation that will pay for extended hospital stays, a little home care, and most drugs. As the law is written, the elderly will pay for the catastrophic illness program themselves with additional premiums on their Medicare insurance. But if inflation in medical costs continues at its 6.6% pace, the government might eventually have to cover a large share of the expenses. Feldstein would also cut the surplus by leaving the taxes that retirees pay on their Social Security income in general revenues. Since 1983 retirees with total incomes over $25,000 ($32,000 for couples) have paid taxes on up to half their monthly benefits. Those revenues, almost $4 billion this year, are dumped into the trust funds. Feldstein argues properly that they should be mixed in with other income tax revenues. Because the $25,000 threshold does not rise with inflation, more and more beneficiaries will be paying the tax as time goes on. By 2000 those taxes would contribute at least $15 billion a year to revenues. These fixes would reduce the annual Social Security surpluses, by $30 billion this year ($26 billion in added costs for doctors' bills; $4 billion in lost tax revenues) and by increasing amounts as the years go on. The goal of the $1 trillion reserve might drift out of reach. So what to do? Congress should pump up the surpluses by slowing down the growth of benefits. Most critics who think Social Security payouts are too ; generous focus on the cost-of-living allowances. True, those are expensive. The consumer price index has been rising a little over 4.5% this year, so inflation alone could add $9 billion to next year's benefits. But current retirees have enormous political power, which makes COLA trimming tough. Slowing the scheduled increases in benefits for future retirees might be more practical than scaling down the inflation protection for today's. ''Initial benefits for new retirees are set too high,'' says Penner, ''and they keep going up. The average guy who retires this year is better off than the one who retired in 1980.'' The extraordinarily complex formula that determines the initial benefit for a freshman retiree provides him with a big bonus for all the gains in wages that have taken place during his career. Consider a 65-year-old retiring this year. The Social Security Administration reviews his earnings history for every month of his career up to age 60 -- and upgrades the earnings. Say he made $250 in August 1960. Wages have gone up about four times since then, so he is credited with almost $1,000 in wages for that month. Each month is indexed similarly. If his average indexed pay was $2,000 a month, the new pensioner will get 90% of what is called in Social Security jargon the first ''bend point,'' currently $319. So he would start with $287.10. Then he would receive 32% up to the second bend point of $1,922, which would give him an additional $512.64. Finally, he would get 15% of the amount over $1,922, which would add $11.70. So his starting basic benefit will be $811.44. A 65-year-old retiring this year can receive a maximum of $838 a month, and his working spouse can get a similar benefit based on his or her career earnings. So a man and wife can collect $20,000 a year or so in old age benefits, half or more tax-free. This only seems fair, you may argue, since adjusting the wage base simply compensates the retiree for inflation over the years. The trouble is that it does more than that. It also builds into his retirement pay the increases in real wages that took place over the period -- the growth in wages that comes from increasing productivity. Why should someone have his retirement benefits calculated as if the level of productivity was as high in 1960 as it is in 1988? In fact, it was 40% lower. OVER THE YEARS the dollar amounts of the bend points will keep rising alongside wages. (After a retiree gets his first check he gets no additional boosts out of rising bend points, however, just cost-of-living increases.) Aldona Robbins, a former Treasury economist who is now a private consultant, calculates that on average a man and wife, both college educated, who retired in 1985 started out with $12,538 a year in old age benefits. A comparable couple retiring in 2010, Robbins estimates, can expect $19,792 in 1985 dollars. The Social Security system would save billions by calculating earnings histories and bend points according to price increases (Alternative IIB anticipates 4% inflation) rather than wage growth (an expected 5.4%). The benefits would still be generous. Robbins calculates that the change would add $11 billion a year to the Social Security surplus by 2000, almost $82 billion a year by 2010. The other major correction the system needs, although it won't affect the surpluses in this century, is an increase in the retirement age. When Social Security started paying benefits in 1940, a 65-year-old man could expect on average to live 12 more years. With better health care a 65-year-old today can look forward to 15 additional years, and a man who turns 65 in 2010 can expect more than 16 years. The reforms of 1983 called for the qualifying age for Social Security to creep up: Those who are now 38 or younger will have to wait until they are 67 to receive full benefits. Right idea, but the qualifying age should be pushed up to 70 and at a faster clip. If baby-boomers wait until 70, millions more people will be putting money into the system in 2020 instead of drawing it out. Adjusting the flow of money into and out of the trust funds so exactly that the reserve will reach a precise level in 2030 is impossible. No policymaker is visionary enough for that, any more than President Roosevelt could have foreseen the conditions of 1988 when he signed the Social Security Act in the midst of the Depression. For now it's enough to start on a $1 trillion reserve and, most important, protect it from plunder.

BOX: FORTUNE'S PLAN

Here's how to build a Social Security reserve that is not too big and not too small, and won't run dry:

-- Reduce the growing surpluses by:

Using payroll taxes to pay all Medicare expenses.

Diverting the income taxes on Social Security benefits into general revenues.

-- Contain benefits by:

Changing the formula for figuring the initial payout that future retirees receive.

Increasing the retirement age to 70.

Bottom line: A reserve of $1 trillion -- just about right -- by 2010.

CHART: NOT AVAILABLE CREDIT: OASDI TRUST FUNDS ILLUSTRATIONS BY MIN JAE HONG CAPTION: SOCIAL SECURITY TRUST FUND DESCRIPTION: Projections for total reserve, outgo and income in social security trust fund from 1988 to 2048; color illustration: working and retired people.

CHART: NOT AVAILABLE CREDIT: NO CREDIT ILLUSTRATIONS BY MIN JAE HONG CAPTION: THE BUDGET DEFICIT IS GREATER THAN YOU THINK DESCRIPTION: Projections for total real budget deficit, Social Security surplus, reported budget deficit, from 1987 through 1993; Two color illustrations of people.