How Social Security works and the promise built into its 1935 design

Pull out a pay stub and find the line marked "Social Security" or "OASDI." Take a hypothetical worker who earns $50,000 in a year, below the annual wage cap. The tax on her wages is 6.2 percent, so 50,000 times 0.062 is $3,100. Her employer pays a matching 6.2 percent, another $3,100. The total sent to Washington on her behalf is $6,200 for the year, a rate of 12.4 percent. A self-employed person pays both halves.

Now the question that matters. Where does that $6,200 go? If you answered "into an account with her name on it, growing until she retires," you have the picture most Americans carry in their heads, and it is wrong. It is the most widespread misunderstanding in American finance, and it has shaped our politics for ninety years. Let me clear it up, because you cannot have an honest debate about Social Security until you know what it actually is.

What the program is

Social Security is a pay-as-you-go system. The payroll taxes collected from today's workers are used, in the main, to pay today's beneficiaries. When she retires, her benefit will be paid mostly from the taxes of the workers who are on the job then. The system is a transfer from one generation to the next, with each generation promised that the next will do the same.

A tiny hypothetical shows the mechanics. Picture an economy with ten workers who each pay $6,200 a year in combined payroll tax. That is $62,000. Suppose it supports two retirees, who each receive $31,000. Now suppose the same ten workers must support four retirees. With the same $62,000, each retiree gets $15,500. To keep each retiree at $31,000, each worker would have to pay $12,400. The ratio of workers to beneficiaries is the whole story. Nothing in that arithmetic is partisan. It is a population fact, and the numbers that drive it are explained in the annual report I describe below.

The 1935 design

On August 14, 1935, President Franklin Roosevelt signed the Social Security Act. It did several things at once. It created a federal old-age insurance program for workers, set at age 65 for full retirement. It set up a federal-state unemployment insurance system. It provided grants to the states for aid to dependent children and for the elderly poor and the blind. The part we call Social Security today is the old-age program, with survivor and disability benefits added later.

The design had choices built into it, and those choices are the "promise" of this essay's title.

1. It was financed by a dedicated payroll tax, not general revenue. The tax began in 1937, at 1 percent on workers and 1 percent on employers, on the first $3,000 of wages. Historians record that Roosevelt and his advisers chose this route partly for political reasons: workers who contribute from their wages feel they have earned the benefit, which makes the program hard to cut.
2. Benefits were tied to earnings, not need. There was no means test for the retirement benefit. A worker with a long, higher-paid career received more, and the formula was weighted to help lower earners. That is why people think of Social Security as something they earned rather than as welfare.
3. It began as a partly funded system. The original plan envisioned building a large reserve. The amendments of 1939 changed course, adding benefits for survivors and the elderly spouses of retired workers and moving the first monthly payments up to 1940, and they moved the program toward pay-as-you-go. The first monthly benefits were paid in January 1940.

The Supreme Court upheld the framework in 1937. In Helvering v. Davis, Justice Benjamin Cardozo wrote for the Court that Congress could tax and spend for the general welfare under Article I, Section 8, and that the old-age provisions fit within that power. So the constitutional foundation is settled. Whether the power ought to be used in a given way is a separate question, and one conservatives have always asked.

The trust funds, and what is in them

Here is the part that confuses nearly everyone.

When Social Security collects more in payroll tax than it pays out in benefits in a year, the law says the surplus must be invested in special Treasury securities and credited to the trust funds. There are two: one for old-age and survivors benefits, and one for disability insurance. They are often analyzed together.

Those securities are not marketable. They cannot be sold on the open market. Each is an IOU from the Treasury to the trust fund and earns interest. When the surplus arrives, the cash goes into the Treasury's general account, where it is spent like any other tax money, on defense, roads or interest. The trust fund holds the bond. When the program later needs cash to pay benefits, the Treasury has to find it by collecting taxes, borrowing from the public, or cutting other spending.

So are the trust fund bonds real? In law, yes. They are legal claims on the Treasury, and Congress would have to honor them. In economic terms, they are a claim of one part of the government on another. They represent a promise that future taxpayers will pay. This is the distinction between the trust fund and a private account: nothing has been saved in a vault. The debt side of the story is told in twelve questions about the national debt, answered straight.

The law also limits what the program can do. Benefits can be paid only from the trust fund's assets and its incoming revenue. If the funds ran out, the program could not borrow to cover the gap. It would be limited to what current payroll taxes bring in, which would be less than the full scheduled benefits. How soon that could happen, and by how much, is something the trustees estimate every year.

What the Supreme Court said

Now the passage that astonishes people.

In 1960, in Flemming v. Nestor, the Supreme Court decided that Social Security recipients do not have a contractual right to their benefits. The case involved a man who had come to the United States as an immigrant, worked and paid payroll taxes for years, and was deported after a 1954 law made past membership in the Communist Party grounds for deportation and cut off benefits for people deported on those grounds. His benefits were terminated. He sued, saying he had earned them.

The Court, in an opinion by Justice John Marshall Harlan, held in a closely divided decision that his claim failed. Benefits, it said, are not accrued property rights protected by the Due Process Clause. The Social Security Act itself, in a section that reserves to Congress the right to alter, amend or repeal any provision, had told everyone from the start that the rules could change. A worker's tax payments, the Court reasoned, do not create a contractual claim in the way that buying an annuity from an insurance company does.

I want to be careful here, because the case is easy to caricature. The Court did not say the government may do whatever it likes. It said Congress has broad authority to change the program, subject to the Constitution's limits, and that the payroll tax is a tax, not a premium. The practical consequences are enormous. Congress can alter benefits, tax them, change the retirement age or modify the formula, as it has several times.

The strongest objection

Here is the opposing view at full strength. A worker pays a dedicated tax for forty years. The statute calls the payments "contributions," under a law called the Federal Insurance Contributions Act. The benefits are computed from the worker's own earnings record. The program is called insurance, the money goes into a "trust fund," and politicians for decades have said you "earned" the benefit. It is reasonable, the critic says, to treat the whole thing as a promise, and a promise ought to be kept, in law as well as in rhetoric.

Here is the answer. First, the moral and political promise is strong, and I would not dismiss it. People have planned their lives around it. A Congress that broke faith with retirees would deserve to lose its job, and the voters have shown that they will punish it. But second, the legal promise is limited by design, and the Court said so. That distinction is the whole point: the strongest guarantee of Social Security is not the Constitution. It is that the voters would not tolerate its being broken. A conservative should understand that and say it plainly, not hide behind the language of an account that does not exist.

Third, the honest conclusion is that the promise to current and near-term retirees should be kept, and that reforms, if they come, should be phased in for younger workers who have time to adjust. Which reforms should be considered is the subject of Social Security's repair options laid side by side in one table. In 2005, President George W. Bush proposed voluntary personal accounts as part of reform, and Congress did not act. The history shows how hard it is to change a program that millions of people rely on.

Two slogans to retire

Both sides of the Social Security argument rely on a slogan that does not survive inspection.

"It's my money, I paid for it." Not in the sense of an account. Your taxes paid for someone else's benefits when you paid them. Your own benefit is a claim on the taxes of people who have not yet paid them. The moral claim is real, because you did what the law required for a lifetime. The accounting claim is not.

"It's a Ponzi scheme." It makes a good applause line, and it is unfair. A Ponzi scheme hides what it is doing and collapses when new money stops. Social Security is public, set out in statute, audited and reported on every year, and it rests on the government's power to tax, not on a promoter's ability to recruit. It does have a structural feature that deserves honesty: it works smoothly when the number of workers is large compared with the number of beneficiaries, and it comes under strain when that ratio falls, as the little ten-worker example showed. Say that, and you are telling the truth. Say "Ponzi scheme," and you give the program's defenders an easy way to dismiss you.

The best conservative position is the plain one. Tell workers what the program is. Keep the promise to those who have relied on it. Make the changes early, so that nobody is surprised.

The 1983 rescue

If you want a model of how to fix this program, look at 1983.

By the early 1980s the system was running short. President Ronald Reagan's first attempt, in 1981, to trim early retirement benefits met a fierce, bipartisan backlash and was dropped. In December 1981 Reagan created the National Commission on Social Security Reform, chaired by Alan Greenspan, a future Federal Reserve chairman. It reported in January 1983. The compromise it reached was turned into law as the Social Security Amendments of 1983, signed on April 20, 1983, in a deal between Reagan, a Republican President, and Speaker Thomas P. "Tip" O'Neill, a Democrat who led the House.

The package was a series of trade-offs, not a miracle. It gradually raised the full retirement age, eventually to 67 for people born in 1960 or later. It made up to half of benefits taxable income for higher-earning recipients, with the proceeds returning to the trust funds. It accelerated scheduled payroll tax increases. It delayed a cost-of-living adjustment by six months. And it brought newly hired federal employees into the system.

The lesson is that both parties gave something up. Neither could have passed it alone. That is the spirit of divided government at its best, the kind described in divided government gave us welfare reform and a balanced budget, and it is a reminder that Reagan, the most conservative President of the era, was willing to sign a deal that raised revenue and trimmed benefits for the sake of keeping the system whole. The story behind the man is in Seven myths about Reagan's economy that both sides still repeat.

The trustees report

Every year the program publishes a check on itself. The Board of Trustees of Social Security and Medicare issues a report that, by law, is due April 1, though it often arrives later. The board has six members: the Treasury Secretary, who serves as managing trustee, the Labor Secretary, the Health and Human Services Secretary, the Commissioner of Social Security, and two public trustees appointed by the President and confirmed by the Senate, who are not government officials in the usual sense.

The report projects the program's income and costs over 75 years. It estimates when the trust funds will be depleted under its intermediate assumptions, and what share of scheduled benefits could still be paid from incoming payroll taxes afterward. I am deliberately not quoting those figures, because they change every year. The report is posted on the Social Security Administration's website, and anyone can read the summary in an hour. When a politician tells you what the program can or cannot afford, open the report and check.

The bottom line

• Social Security is a pay-as-you-go transfer. Your payroll tax pays today's retirees, and tomorrow's workers will pay yours. It is not a savings account.
• The trust funds hold special Treasury securities. They are legal claims on the Treasury, but not a pile of saved cash.
• The Supreme Court held in Flemming v. Nestor in 1960 that there is no contractual right to benefits. The real guarantee is political, which means it depends on you.
• The 1983 rescue by Reagan and O'Neill shows that this program can be repaired when both parties give something up.
• Read the latest trustees report on the Social Security Administration's website, then ask your members of Congress how they would keep the promise to retirees and be honest with workers. See also a paycheck worksheet that shows where every dollar of your pay goes.