A Flimsy Trust: Why Social Security Needs Some Major Repairs

(By Allan Sloan, August 2, 2009)

In Washington these days, the only topics of discussion seem to be how many trillions of dollars to throw at health care and the recession, and whom on Wall Street to pillory next. But watch out. Lurking just below the surface is a bailout candidate that may soon emerge like the great white shark in "Jaws" -- Social Security.  Perhaps as early as this year, Social Security, which at $680 billion is the nation's biggest social program, will be transformed from an operation that's helped finance the rest of the government for 25 years into a cash drain that will need money from the Treasury. In other words, a bailout.

I've been writing about Social Security's problems for more than a decade, arguing that having the government borrow several trillion dollars to bail out the program so it can pay its promised benefits would impose an intolerable burden on our public finances. But I've changed my mind about what "intolerable" means. With the government spending untold trillions to bail out incompetent banks and the auto industry, it should damn well bail out Social Security recipients, too. But in a smart way.

Why am I talking about Social Security now, when health care is sucking up nearly all the oxygen in our nation's capital? Because Social Security is a big deal, providing a majority of the income for more than half of Americans 65 and up and also supporting millions of people with disabilities and survivors of deceased workers. And because the collapse of stock prices and home values makes Social Security retirement benefits far more important than they were during the highs of a few years ago. And because the problems aren't that hard to solve if we look at Social Security realistically instead of treating it as a sacred, untouchable program (liberals) or a demonic plot to make people dependent on government (conservatives).

Finally, this is a good time to discuss Social Security because the Obama folks say it's next on the agenda, after health care. No one at the White House, the Treasury Department or the Social Security Administration would discuss specifics, however. It ought to tell you something that Peter Orszag, director of the White House Office of Management and Budget, is a noted Social Security scholar. Alas, he wouldn't tell me what he plans to propose. "Health care first" was all he'd say.  I'd like to show you that Social Security has a real and growing cash problem even as its trust fund is getting bigger than ever, explain how the program really works, and -- immodest though it may seem -- propose a few solutions.

The Cash Problem

How can Social Security possibly need a bailout when, by Washington rules, it's "solvent" for another 26 years? To understand the problem, look at me. I'll turn 66 next year, which makes me and my wife eligible for full Social Security benefits. They'll be about $42,000 a year for the both of us starting Jan. 1, 2011, and are scheduled to rise as the consumer price index does.  Social Security, which analyzed my situation, values those promised (but not legally binding) benefits at a bit more than $600,000. That is a lot of money, but Social Security is way ahead of us because the value of our benefits is far less than the Social Security taxes we and our employers will have paid by the end of next year, plus the interest Social Security will have earned on that money in the decades since we started working. Those taxes and interest will total more than $800,000 by Dec. 31, 2010. For example, the $5.18 my employer and I paid in 1961 -- the year I got my card -- will have grown to $140 by next year.

I don't have a problem with this disparity. One of the principles of Social Security is that higher-paid folks like me support the lower-paid. That's as it should be, given that the Social Security tax (12.4 percent of covered wages, split equally between employer and employee) is regressive, far more costly as a percentage of income to a $40,000-a-year worker than it is to me. According to the Tax Policy Institute, five of six U.S. workers pay more in Social Security tax (including the employer's portion) than in federal income tax -- something that makes it especially important (and only fair) to preserve the program for lower earners, who get old-age benefits of up to 90 percent of their covered wages, while I get only 28 percent.

How can my wife and I pose a problem to Social Security when our benefits are valued at $600,293, while our tax payments plus interest will total $804,686? Answer: Because the obligation is real, but the $800,000-plus asset is illusory, consisting solely of government IOUs to itself.  Now, let's step back a bit -- to 1935, actually -- to see how we got into this mess. President Franklin D. Roosevelt set up Social Security as an intergenerational social-insurance plan, under which today's workers support their parents (and those with disabilities and workers' survivors) in the hope that their children will in turn support them. It's not a pension fund. It's not an insurance company.

Social Security exists in its own world. In this world, taxes are called "contributions," though they're certainly not voluntary. "Trust funds," which in the outside world connote real wealth bestowed on beneficiaries, are nothing but IOUs from one arm of the government (the Treasury) to another (the Social Security Administration). And "solvency," which in the real world means that assets are greater than liabilities, means only that the Social Security trust fund has a positive balance.  Alas, the trust fund is a mere accounting entry, albeit one with a moral and political claim on taxpayers. It currently holds about $2.5 trillion in Treasury securities and is projected to grow to more than $4 trillion, even as Social Security begins to take in far less cash in taxes than it spends in benefits. For instance, it projects a cash deficit of $234 billion for 2023. But the trust fund will grow -- on paper -- because it will get $245 billion in Treasury IOUs as interest. The Treasury pays its interest tab with paper, not cash.  "The trust fund has no financial significance," says David Walker, former head of the Government Accountability Office. "If you did [bookkeeping like] that in the private sector, you'd go to jail."

Let me show you why the Social Security trust fund isn't social or secure, has no funds, and can't be trusted, by returning to my favorite subject: myself.  The cash that Social Security has collected from me and my wife and our employers isn't sitting at Social Security. It's gone. Some went to pay benefits, some to fund the rest of the government. Since 1983, when it suffered a cash crisis, Social Security has been collecting more in taxes each year than it has paid out in benefits. It has used the excess to buy the Treasury securities that go into the trust fund, reducing the Treasury's need to raise money from investors. What happens if Social Security takes in less cash than it needs to pay benefits? Watch.

Let's say that late next year, Social Security realizes that it's short the $3,486 it needs to pay me and my wife for our Jan. 1, 2011, benefit. It gets that money by having the Treasury redeem $3,486 in trust-fund Treasury securities. The Treasury would get the necessary cash by selling $3,486 in new Treasury securities to investors. That means that $3,486 has been moved from the national debt that the government owes itself, which almost no one cares about, to the national debt it owes investors, which almost everyone -- and certainly the bond market -- takes very seriously.

This example shows you that the trust fund is of no economic value to the government as a whole (which is what really matters), because the government has to borrow from private investors the money it needs to redeem the securities. It would be the same if the trust fund sold its Treasury securities directly to investors -- the government would be adding to the publicly held national debt to fund Social Security checks. Social Security's "solvency" calculations -- and the insistence by the status quo's supporters that there's "no problem" until 2036 because the trust fund will have assets until then -- assumes that the Treasury can and will borrow the necessary money to redeem the trust fund's Treasury securities. There is also the assumption that our children, who by then will be running the country, will allow all this money to be diverted from other needs. I sure wouldn't assume that.

This whole problem of Social Security posting huge surpluses for years, using proceeds from a regressive tax to fund the rest of the government and then needing a Treasury bailout to pay its bills, is an unanticipated consequence of the 1983 legislation that supposedly fixed the system.  In order to show 75 years of "solvency" as required by law, Congress, using the bipartisan 1983 Greenspan Commission report as political cover, sharply raised Social Security taxes, cut future benefits and boosted the retirement age (then 65, currently 66, rising to 67).

The changes transformed Social Security from an explicitly pay-as-you-go program into one that produced huge cash surpluses for years followed by huge cash deficits. No one in authority seems to have realized that the only way to really save the temporary surpluses was to let the trust fund invest in non-Treasury debt securities, such as high-grade mortgages (yes, such things exist) or corporate bonds. That way, interest and principal repayments from homeowners and corporations would have been covering Social Security's future cash shortfalls, rather than the Treasury's having to borrow money to cover them.

This problem has been metastasizing for 25 years. Now I'll show you why the day of reckoning may finally be here.  Just last year, Social Security was projecting a cash surplus of $87 billion this year and $88 billion next year. These were to be the peak cash-generating years, followed by a cash-flow decline, followed by cash outlays exceeding inflows starting in 2017.  But in this year's Social Security trustees report, the cash flow projections for 2009 and 2010 have shrunk by almost 80 percent, to $19 billion and $18 billion, respectively. How did $138 billion of projected cash go missing in one year? Stephen Goss, Social Security's chief actuary, says the major reason is that the recession has cost millions of jobs, reducing Social Security's tax income below projections.

But $18 billion is still a surplus. So why do I say Social Security could go cash-negative this year? Because unemployment is far worse than Social Security projected. It assumed that unemployment would rise gradually this year and peak at 9 percent in 2010. Now, of course, the rate is 9.5 percent and rising -- and we're still in 2009.  Social Security's having negative cash flow this year would be a relatively minor economic event -- what's a few more billion dollars when the government's already borrowing more than $1 trillion? -- but I think it would be a really important psychological and political event.

Orszag pooh-poohed my thinking when I met with him. He says I'm wrong to harp on Social Security's near-term cash flow -- a term, by the way, that he won't use. "I think the real question of Social Security is how we bring long-term revenues in line with long-term expenses," he said, "not whether the primary surplus within Social Security turns negative within the next few years." I guess we'll see.  When you look back at numbers from previous years, you suddenly realize that Social Security's finances have been deteriorating for a long time. Social Security's cash flow (and thus its trust fund balances) has fallen well below earlier projections. Seven years ago, the projected 2009 cash flow was $115 billion. That fell to $87 billion by last year and is now $19 billion. Ten years ago, the trust fund was projected to be $3 trillion at the end of this year, rather than the currently projected $2.56 trillion.

In 1983, the system was projected to be "solvent" until the 2050s. This year it's only until 2036. Social Security's Goss says the major reason is that over the past two decades, the wages on which Social Security collects taxes have grown more slowly than projected. He said Social Security projected them to grow at 1.5 percent above inflation, but they've been growing at only 1.1 percent above it.  The scariest thing, at least to me, is that even as its financials erode, Social Security is as important as ever -- maybe more so. Let me elaborate on what I said earlier, about how older people depend heavily on Social Security. It accounts for more than half the income of 52 percent of married couples over 65, and 72 percent of that of 65-and-up singles, according to the Social Security Administration. What's more, this dependence -- which Goss says isn't projected to change -- comes despite 30 years of broadly popular self-directed retirement accounts such as 401(k)s, IRAs, 403(b)s and such.

Why haven't those savings accounts reduced dependence on Social Security? Part of the reason is that it takes a lot of money to generate serious retirement income: about $170,000 for a $1,000-a-month lifetime annuity. Inflation protection, if you can find it, is ultra-expensive. Vanguard, which offers a lifetime inflation-adjusted annuity in conjunction with an AIG insurance company called American General, quoted me a staggering price for an annuity mimicking my wife's and my Social Security benefit. Would you believe $774,895?  Another problem is that the stock market has been stinko. Stocks are below their level of April 2000, when the great bull market (August 1982 to March 2000) ended. It's hard to make money in stocks when they've been down for nine years. The Employee Benefit Research Institute estimates that the average retirement account balance of people 65 to 74 was $266,000 in 2007 but had fallen to $217,000 as of mid-June.

Then there's the problem of lost home equity. According to a study conducted for Fortune by the Center for Economic and Policy Research, people in the lower-income to upper-middle-income ranges have lost a far greater proportion of their net worth as a result of the housing bust than the most wealthy people have.  The bottom line is that many older people who felt reasonably well fixed for retirement a few years ago now need Social Security more than ever. That makes it even more important to come up with a way to sustain it and to show our children a realistic plan to give them benefits, rather than to rely on the trust fund and the supposed political clout of the geezer class to keep benefits flowing when cash flow goes negative. So how do we fix these problems? Let me divide it into three categories: what to do, what to change and what not to do.

What to Do

Many of the old standbys: raising the "covered wage" limit, but not to outrageous levels; tweaking the benefit formulas so that high-end people like me get a little less bang for the buck; modifying cost-of-living increases for us high-end types; and, most important, raising the retirement age to 70, with a special earlier-retirement provision for manual laborers, who can't be expected to work that long.  The law requiring 75-year solvency. It's hard to predict what will happen 75 days from now, let alone 75 years from now. But the obsession with 75-year solvency and the status of the trust fund has obscured what's really going on. This requirement forces Social Security's actuaries -- who are among the best and smartest public servants I know -- to make all sorts of impossible projections. As we've seen, even one faulty projection -- such as overestimating wage growth -- can cause substantial problems.

The trust fund. Before the Greenspan Commission-related changes in 1983, the trust fund was a checking account. The workings of Social Security since 1983 have turned it into something it was never intended to be: an investment account. Let's gradually draw down the trust fund by having the Treasury redeem $100 billion or so annually (less than the current interest the fund earns) by giving the fund cash rather than Treasury IOUs, gradually increasing the redemptions. That will let the fund buy assets that will be useful when serious cash-flow deficits hit, assets such as high-grade mortgage securities and high-grade bonds.  That way we'll be bailing out Social Security a bit at a time, which is realistic, rather than in huge chunks, which isn't. Combine that with the lower costs and higher revenues, and today's kids could see that there really is a way they'll get benefits someday.

What Not to Do

Depend on taxing "the rich." One solution you hear in Washington is restoring "covered wage" levels to the good old Greenspan Commission days, when 90 percent of wages were subject to Social Security tax, compared with 83 percent now. Sounds simple and fair, doesn't it? But that would increase the Social Security wage base to about $170,000 from the current $106,800, according to Andrew Biggs of the American Enterprise Institute -- at 12.4 percent, a huge new tax to middle-class workers. (And yes, that's middle-class income, not rich-person income, in large parts of the country.)

During his campaign, President Obama proposed (and then dropped) a plan to leave the Social Security wage cap where it is but to apply the 12.4 percent Social Security tax to all wages above $250,000. That -- like the 90-percent-level-of-income idea -- would be a huge new tax that would weaken support for Social Security among higher-income people. I'm not saying "rich people," because truly rich people generally have huge amounts of investment income, which isn't subject to Social Security tax. Means-test benefits. It's being done. We'd be making a terrible mistake to means-test Social Security by saying that people above a certain income level can't get it. That would violate the social compact that everyone pays Social Security taxes and everyone gets something.

Besides, Social Security is already means-tested, indirectly. That's because if you have enough non-Social Security income -- about $23,000 a year in my case -- you pay federal income tax on 85 percent of your benefit.  Given the three pensions I stand to collect from previous employers, I think I hit that level. So, for the final time, let's run my numbers. If my wife and I are in the 28 percent federal tax bracket when we start collecting benefits, we'll be giving almost a quarter of our benefit right back to Social Security. It would also mean that the $600,000 benefit I talked about earlier would cost Social Security only about $450,000 -- just 55 percent or so of the $800,000-plus value of our taxes.

I don't mind that big haircut, but I'd be furious if the government decided to just confiscate all the money my wife and I put in over the decades by saying we were "rich" and had no right to any benefits. And I wouldn't be alone.  Given the way health-care reform has bogged down, Social Security may not make it onto the agenda until next year. But it's going to show up sooner or later, probably sooner, because the numbers are so bad that something's going to have to be done. As I hope I've shown, we're going to have to bail out Social Security or risk hurting a lot of low-income older people or putting the whole program at risk by gouging and alienating upper-income Social Security sympathizers like me.  So let's fix this already. By the numbers. And by the right numbers, not fantasy ones.