An idea whose time has come: the necessity and desirability of Social Security reform.

PEOPLE think that Social Security reform is an incredibly complicated and difficult subject. But it's not, really. The program's benefits are rising faster than its revenues, which means we're going to have a lot of trouble paying for those benefits. The program also offers young people a lousy deal. The solution is, first, to keep future benefits from growing so fast. But acknowledging that benefits have to shrink doesn't make the program a better deal for young workers. So, second, something should be done that lets them build up retirement savings to make up for it. Which means their taxes should be cut a bit, too, on the condition that they invest the money they get for retirement. That, in a nutshell, is the reform President Bush seems inclined to propose. There: Was that so hard?

It's not the reform itself, but the politics of reform that is difficult, and may well sink Bush's proposal in Congress. And it's politics that has generated much of the confusion over basic facts in the Social Security debate.

Most Democrats already oppose Bush's reform, and they have many reasons for opposing it--starting, in many cases, with the genuine conviction that it is a bad idea. But it is very hard to have respect for people who are trying to discredit reform by minimizing the program's problems.

Social Security has mostly been a "pay as you go" program. Today's workers finance, with their payroll taxes, the retirements of today's senior citizens. When those workers get old, they will in turn rely on taxes from their children's generation. But increased lifespans and decreased childbearing have made this bargain hard to sustain. In 1950, there were 16 workers supporting each retiree. There are 3.3 today. By 2040, there will be only 2.1.

Benefits are calculated using a complex formula. If you had higher wages, and thus paid more payroll taxes into the system, you'll get a higher benefit. But benefits don't rise in exact proportion to wages: There's some (somewhat haphazard) redistribution to the poor. Homemakers get a benefit, too; the system, it might be said, recognizes that they have contributed to it by nurturing the next generation of taxpayers. During your retirement, your benefits go up every year to keep up with inflation.

A TRILLION HERE, A TRILLION THERE

A worker who has earned the average American worker's wage each year will get $14,854 from Social Security this year if he retires now. Because wages grow over time, so do benefits. By 2050, the average worker's retirement benefit is expected to be $23,811 (adjusting for inflation). That's 60 percent higher than today.

That growth is what is causing the fiscal problem. In 2018, the program is expected to start sending out more checks than it collects in revenues. The system has an expected $283 billion shortfall during the following five years. By 2045, it is expected to lose $394 billion in one year. (There's an asterisk by that number, though, as we'll see below.) One way to measure the financing gap is to estimate how much additional money the government would have to devote to Social Security now to make the program permanently solvent. Depending on how you calculate it, that number is $10 to 11 trillion. That's larger than our entire economy.

The president says that "the crisis is now." That comment has inspired a lot of fairly tedious semantic debate. Let's just say that we have a serious problem. It is true that we do not have to fix it immediately. It is also true that every year we wait, the choices get worse. We can gradually cut benefits if we get started now. If we don't, we will have to cut them (or raise taxes) very sharply. Andrew Biggs, now a commissioner at the Social Security Administration, has estimated that delaying reform for one more year will cost $600 billion--and that cost goes up every year. The Titanic didn't have a crisis until it hit the iceberg, but it would have been better off gently steering a different course beforehand.

Liberals contend that the scenario I have painted above is alarmist. Social Security, they say, will not actually go bankrupt until 2042. The program can be saved with some minor adjustments. Just rolling back Bush's tax cuts would raise the necessary funds. If the economy grows better than expected, the program might have enough revenues to pay for its promises.

None of this is true. The idea that the problem does not start until 2042 depends on sleight of hand. For several decades, the program, in anticipation of the retirement of the Baby Boomers, has collected more revenues than it pays out. The surplus has been banked in a Social Security trust fund. The liberal argument is that when payouts start to outstrip revenues, in 2018, the program can just draw on the trust fund--and it can keep drawing on it until 2042, when it is scheduled to run out.

Liberals--I mean you, Paul Krugman--have spent immense amounts of verbiage obscuring the fact that the trust fund is an accounting fiction. Its assets are IOUs from the rest of the federal government. When 2018 rolls around, the government will have to find the money to pay off those IOUs. It will have to raise revenues or cut spending or borrow elsewhere to do that.

FORMS OF DENIAL

The "minor adjustments," meanwhile, aren't all that minor. The leading liberal plan on Social Security is that of economists Peter Diamond and Peter Orszag. Its backers, such as Jonathan Cohn of The New Republic, prefer to describe it as "a series of tweaks" that "involves both raising taxes and cutting benefits." Likewise, the Washington Post describes Diamond and Orszag as "'balancers,' who would use benefit cuts and tax increases in equal measure." In truth, 85 percent of the improvement in solvency their plan accomplishes would come from higher taxes. (They even propose adding some benefits.) They want to raise the payroll-tax rate. They want to raise the amount of wages to which the tax applies. Then they want to add a "surtax" to high earners.

The Congressional Budget Office (CBO) estimates that the Diamond-Orszag plan would mean that the share of the economy eaten up by Social Security taxes would grow by more than a third. (Those taxes equal 4.9 percent of the economy now but would rise to 6.6 percent by 2080 under the plan.) The cost of their plan over the next 75 years would be almost exactly the same as that of the current program. The CBO also estimates that the plan would reduce economic growth, the capital stock, and work effort.

The idea that merely repealing the Bush tax cut would pay for Social Security's promises is misleading. The people who make that claim are, again, ignoring the problem of paying off the IOUs in the trust fund. (They're assuming, that is, that the trust fund can pay for Social Security benefits painlessly.) And there's another problem. Because the income tax is progressive, its burden grows over time. As wages rise, people move into higher tax brackets. The average tax rate--the share of the economy that the government is claiming through taxes--therefore rises. The Alternative Minimum Tax exaggerates this trend.

In a recent issue of Tax Notes, Brian Jenn and Donald Marron note that current law would have the federal government taking an unprecedented 20.9 percent of the economy in 2018 and 27 percent by 2078. That's just by going on autopilot. The people who say we could pay for Social Security by repealing Bush's tax cuts are assuming taxes would go even higher than that. It's an unrealistic, and undesirable, assumption.

Finally, there is the illusion that we can grow our way out of the Social Security shortfall. One of the business economists at Bush's own economic summit lent credence to this view. But the flaw is pretty obvious. Sure, higher growth would yield revenues. But higher growth would also increase wages--and since benefits are tied to wages, benefits would rise too. Growth is a very good thing, but it can't relieve us of the need to get the program's benefits structure in line with its revenues. Even if economic growth doubled, Social Security would still go insolvent. And to the degree that economic growth does help, it is mostly by swelling the balance of the trust fund, giving the system more IOUs to claim against future taxpayers.

CHECKING THE INDEX

Hence Bush's apparent support for changing the formula by which benefits are calculated. He wants benefits to keep pace with inflation--but not to grow any faster than that. The middleincome worker of 2050 wouldn't get a benefit that's 60 percent higher than that of the middle-income worker of today: He'd get the same benefit. That means that when people retire, Social Security would be replacing a smaller portion of their paychecks.

Moving from "wage indexing" to "price indexing" isn't a minor change. The middle-income worker of 2050 would be getting an annual benefit worth 37.5 percent less than he would have gotten under wage indexing. Price indexing would eliminate Social Security's shortfall all by itself. Would it be a draconian cut in benefits? If wages grow over time, workers will be putting more tax money into Social Security: Shouldn't they get bigger benefits as a result?

What that question ignores is that Social Security is not capable of converting our worker's payments into those massively higher benefits. The only way he could get those benefits is if he agreed to pay more taxes over the course of his working life (and to work in the smaller economy caused by higher taxes on everyone). The system can't pay for the larger benefits without tax increases. So nothing he could actually have, on terms he would want to have, would be taken away from him. If our worker wanted more money for retirement, he would almost certainly prefer to invest additional money himself rather than have the government raise his taxes.

In the debate over Social Security reform, even knowledgeable reformers are being careless in using phrases like "promised benefits" and "current-law benefits" to describe the 60 percent increase our median worker might get under wage indexing. The Supreme Court has ruled that nobody has a legal right to a particular benefit level. And current law includes a provision that automatically slashes benefits when the "trust fund" runs out in 2042. The current projections have benefits getting cut by 27 percent--and not gradually, but in one year, with more cuts following. (And this is after the program has already generated $5.8 trillion worth of debt as the trust-fund IOUs are redeemed.) If you take account of that, Bush's cuts don't look so big after all.

THE CASE FOR PERSONAL ACCOUNTS

If price indexing will balance the books, it might be asked, why does Bush also want to introduce personal accounts? There are three answers to that question.

The first is that solvency isn't the only problem with Social Security's current structure. It takes a lot of money from young workers without giving them much in return. The program can afford to give a middle-income 25-year-old only 91 cents for every dollar he is going to put in over the course of his working life. Price indexing recognizes that reality, but does nothing to improve it. Personal accounts are a way of softening the blow. People would be able to supplement their reduced Social Security checks with the wealth built up in their accounts.

The second way to look at it is that the accounts are a way of putting Social Security on a firm footing. We would be moving from a "pay as you go" model in which each generation funds the retirement of the next to a "pre-funding model" where each generation saves in advance for its own retirement. If the government tries to do the pre-funding for us, without letting us have the money in individual accounts, there's a greater risk that the money will be diverted to other programs. The government could also gain too much influence over the economy if it tries to invest all that money itself.

The third view is that making it possible for millions of Americans to buy a share of the economy through personal accounts would be a good thing in itself. It would give them wealth they could hand on to their heirs. It would make them worker-capitalists with a direct interest in sound economic policies. Edward Prescott, the 2004 Nobel Prize winner in economics, argues that letting people invest payroll taxes would also expand the economy by stimulating them to work more. The CBO has found that a plan including price indexing and personal accounts would lead to a larger economy. National wealth would be 10 to 12 percent larger by 2080. (That's in contrast with the wealth-destroying Diamond-Orszag plan.)

Liberal politicians and journalists persist in speaking of capital-market investment as though it were "speculation" or "roulette." But nobody is talking about having 64-year-olds take all of their Social Security money and put it all on a tech stock. We are talking about letting people invest a portion of their Social Security contributions over a long period of time. There are plenty of ways to reduce the risks--by diversifying stock portfolios, by buying index funds, by buying bonds as well as stocks. People are quite capable of taking advantage of these methods (and a law establishing personal accounts would probably require them to do so). Between the end of 1999 and the end of 2002, broad stock-market indices fell by nearly 40 percent. But a report jointly produced by the Employee Benefit Research Institute and the Investment Company Institute found that the average 401(k) plan fell by only 10 percent. It's established fact that stocks grow less volatile over the long run: There is no 20-year period in American history in which the stock market has declined. The average annual return on stocks has been 6.7 percent above inflation.

Setting up personal accounts does, however, involve a short-term increase in the federal debt. This debt increase is the badly misunderstood "transition cost" to personal accounts. Since we have a pay-as-you-go program, most of today's payroll taxes go to today's retirees. The rest goes, notionally, to the trust fund, but really to fund other government operations (which borrow from the trust fund). If people can put some of their payroll taxes in personal accounts, the government has to find other money to fund its other commitments. The result could be $1 trillion or even $2 trillion of borrowing over the next decade.

Those numbers may look alarming. But it's important to remember that they do not represent new costs. The government is already undertaking to pay for the retirements of today's senior citizens and tomorrow's. Pre-funding part of tomorrow's retirements would not increase the government's total obligations; it would just bring some of them forward in time. If the accounts are part of a long-term deal to reduce the cost of Social Security--as in Bush's price-indexing plan--then temporary debt increases should not deter us. The markets are well aware that the future benefits from Social Security constitute a large off-budget liability for the government. An increase in on-budget debt, even a large one, will not frighten them if they have confidence in the long-term trajectory of the reform plan.

DECISIONS, DECISIONS

Republicans, and Bush especially, face a daunting series of questions as they figure out how to move forward on Social Security. How much should they let people invest in personal accounts? Should it vary by income, so that poorer people can invest a higher percentage of their wages? How tightly should the government regulate the investments? Should it require people to convert their accounts into annuities when they retire? Should Republicans take the political risk of cutting benefits through price indexing? How should they finance the transition? Should the White House devise a plan at all, or let one emerge from Senate horse-trading?

There are trade-offs attending each choice. Large accounts would give workers a reason to participate, but would also mean higher transition costs. Or take the question of annuities. Frank Keating, the head of the American Council of Life Insurers (many of which sell annuities), argues that people shouldn't be able to take their whole account when they retire and blow all the money in Vegas. If they bought annuities with the money, they would have a guarantee that they will have an income stream as long as they live. That way, they won't end up destitute--and clamoring for Washington to bail them out. Regulations to keep people from impoverishing themselves would help get a reform bill through Congress.

On the other hand, if the whole account has to be made into an annuity, then some of the appeal of personal accounts vanishes. Reformers would not be able to say that the accounts are a way of building wealth that can be handed down the generations. The reformers' argument that Social Security is an especially bad deal for blacks, who have shorter lifespans than whites and thus receive less in benefits, would also lose much of its force. These considerations point to a compromise: Account holders should be required to buy an annuity that keeps them out of poverty, but free to do whatever they want with any additional funds they have accumulated.

There is a political theory that can help guide reformers in making some of these choices--a theory that falls under the heading of "the new investor class." The theory holds, among other things, that participation in capital markets changes people's political attitudes and behavior. Perhaps more pertinently here, it holds that pro- investor policies tend to perfect themselves over time. So, for example, a limited tax break for retirement savings will generate a constituency that wants bigger tax breaks for saving for retirement as well as for health and educational expenses.

A relatively small personal-account option, for the same reason, will generate demands for its expansion. Reformers need not be disappointed if they don't win large accounts this year. Holders of 401(k)s have over time gotten more and more options in how they invest their money. Regulations on personal accounts will probably get looser as well.

But the theory, by itself, is not a substitute for the political judgment required to make the major choices here. The theory is capable of telling conservatives that instituting personal accounts offers the prospect of reshaping American politics in their favor. The upside from the accounts is larger than from policies such as expanded IRAs, which would not do as much to increase the number of investors. But the political risks are greater, too, and the theory cannot tell you whether a specific historical moment is ripe for reform. If reform fails this year, Republicans may rue not having taken a more incremental route.

Congressional Republicans are of two minds about Social Security reform. They know that personal accounts have risen in popularity. They know, too, that Republicans--including Bush--have broached the subject of reform without losing elections in 2000, 2002, and 2004. We have come a long way from 1988, when Jack Kemp denounced Pete du Pont in the presidential primaries for daring to suggest "privatization." But Republicans are also nervous because they don't think there is a mandate for a large debt increase, however temporary, or for price indexing. Unfortunately for them, the leading proposal that avoids price indexing carries a much larger debt with it.

The Democrats have their own divisions. They haven't decided whether to come up with a plan of their own, or to put all their eggs in the there's-no-crisis basket. But what makes it hard to see Bush prevailing is that there isn't a Democrat on Capitol Hill who is afraid to oppose him on this issue. The twelve Senate Democrats who come from states that Bush won by more than 5 points have good reason to think twice before opposing one of his Supreme Court nominees. When it comes to Social Security, however, all the political pressures on them will be to resist reform. Yet Bush will probably need at least five Senate Democrats to beat a filibuster and enact a bill.

House Republicans are not going to vote for price indexing only to have a reform bill die in the Senate. Once a bill passes, they will be able to tell the voters that the sky hasn't fallen. If it doesn't pass, however, they will face ads saying that they "tried to cut Social Security"--and plenty of older voters will think that the Republicans were going after them, not future oldsters. So House Republicans want the Senate to go first. Many of them also want the president to offer a bill to serve as a template for Congressional legislation. Some Senate Republicans are happy to go before the House, but also want to go before the White House. They want to cobble together a bipartisan deal. If the Senate Republicans can recruit enough Democrats to their side, they will be in a strong position to dictate the terms of an eventual deal.

Whatever emerges from Congress, if anything does, will include provisions that conservatives do not like: perhaps an increase in the progressivity of the payroll tax, or something worse. But what Bush appears to have in mind--creating personal accounts and restraining the growth of Social Security--is the right policy. If he manages to get that policy enacted, it will not only be the signal domestic achievement of his presidency. It will be the biggest legislative victory in the modern history of conservatism.