Contrary to conventional thinking, allowing workers to invest a portion of their Social Security taxes through personal retirement accounts will incur no net cost to the federal government. The misconception that there will be new costs–i.e., “transition costs”–arises from forgetting that Social Security is already, in effect, debt-financed. Consequently: 1) Any debt sold to the public in the short term during the so-called “transition period” to pay all promised Social Security benefits simply refinances an existing liability; and 2) Refinancing Social Security’s debt through personal accounts makes possible a reduction, and eventually an elimination, of Social Security’s unfunded liability, which otherwise is scheduled to grow without bound. Mountain Of Debt In Social Security, the current generation of workers pre-funds its retirement benefits by making contributions to the program in the form of a tax taken out of every paycheck. In exchange for this tax, the government promises to pay the taxpayer, when he retires, precisely defined benefits, which are specified in law by formula. The amount of money paid in payroll taxes would be more than adequate to pre-fund workers’ retirement benefits were it properly invested, but it is not. In effect, the federal government borrows the payroll tax payments of current workers to pay the retirement benefits to current retirees. But unlike ordinary government debt, the debt obligation owed to workers through Social Security is not quantified and memorialized by the issuance of bond, note, or bill certificates to the workers. Nor is it formally recorded as debt on the balance sheet of the U.S. government. It is real debt nonetheless. What pundits have labeled the “transition cost,” is really the short-term cash-flow crunch that will happen when workers are allowed to invest a part of their Social Security taxes in personal retirement accounts rather than loan that money to the government to pay the benefits of current retirees. Cash Flow This cash-flow problem, however, does not arise because personal accounts create new (i.e., “transition”) costs. To the contrary, the cash-flow problem arises because the government would be borrowing less. Every dollar of a worker’s Social Security payroll tax contribution invested in a personal retirement account is a dollar less debt the government owes to the worker and would otherwise have to pay back in the form of future Social Security benefits. Congress has a choice of how to manage this temporary cash-flow shortage. It can refinance the old (undocumented) debt obligation to Social Security beneficiaries through some new formal debt instrument, i.e., federal bonds. It can forgo refinancing this debt and actually reduce the country’s current indebtedness by cutting planned future federal spending and use the freed-up revenue to repay the debt to former workers in the form of Social Security benefits. Or it can increase taxes to raise the required cash. Economically, the most rational solution would be to require Social Security personal-retirement-account holders to lend the federal government whatever funds are needed to pay current Social Security benefits that otherwise would have been covered by the personal-retirement-account holder’s payroll taxes. In exchange, the federal government would deposit into the taxpayer’s personal-retirement-account Treasury-Inflation-Protected Securities (TIPS), interest-bearing long-term federal bonds backed by the full faith and credit of the United States with no restrictions on the right of the account holder to resell the bonds in secondary bond markets. Refinancing the Social Security liability in this way would not cause a short-term shock to the bond market, and would not create immediate upward pressure on interest rates, because the government would not be entering credit markets to raise new cash. It would simply issue new bonds. Nor does this approach pose a long-term threat to financial markets since the bond market is already fully aware of the unfunded Social Security liability. The future financial obligation represented by the Social Security liability is already reflected in the current price of federal bonds. Since new bonds issued to personal retirement accounts would simply refinance an already existing liability, no net increase in federal indebtedness results. There are no “transition costs” involved.
Federal Debt Under Current Social Security Law v. Ryan/Sununu Assuming No Revenue Capture
Not only would refinancing the Social Security liability through personal retirement accounts not entail new debt, it would make it possible to pay off that debt and leave Social Security financially sound in perpetuity. The chart above illustrates this fact. The red line represents the total new debt (as a share of GDP) that will be incurred under current law if there are no benefit cuts or tax increases. The green line depicts the amount of borrowing relative to GDP required to capitalize “large” personal retirement accounts (i.e., about half of all Social Security payroll tax contributions) without cutting Social Security benefits, raising taxes or increasing the retirement age, if Congress imposes only modest spending restraint. The amount of spending restraint shown in the chart is based on the spending-control mechanism proposed in the legislation introduced by Rep. Paul Ryan (R.-Wis.) and Sen. John Sununu (R.-N.H.), which is designed to reduce the growth rate of federal spending one percentage point a year below CBO’s baseline for eight years–i.e., 3.6% a year vs. 4.6% projected on average. This would lower total federal spending as a share of GDP in the coming decade from CBO’s projected 20% to about 18.4%–the historic average for federal revenues during the past 40 years and exactly what CBO projects federal revenues would be in the future if all of the Bush tax cuts were made permanent and the AMT were reformed. The chart conservatively–and unrealistically–also assumes no positive dynamic revenue feedback effect from creating the accounts. The area between the two lines to the left of their intersection–usually mistakenly characterized as “transition costs”–is significantly smaller than the “do-nothing costs” represented by the area to the right of the intersection, which increases without bound beyond 2037. Thus, the debt incurred to refinance the Social Security liability (which reaches a maximum of 29% of GDP in 2032) and alleviate the cash crunch when workers are allowed to invest their Social Security payroll tax contributions–so-called “transition costs”–not only isn’t new debt, it is more than offset in later years when personal accounts reduce and eventually eliminate the Social Security liability altogether. ‘National Blessing’ President Bush and Congress have a golden opportunity to modernize Social Security and make it solvent in perpetuity through personal retirement accounts. They should start by allowing workers to place about half of their Social Security payroll tax contributions into personal retirement accounts. Then, rather than panic over fictitious “transition costs,” they should alleviate as much of the cash crunch as they can by restraining federal spending growth and enacting tax reforms to spur faster economic growth and generate as large a dynamic revenue feedback effect as possible. After that, there is no need to worry about the rest. Congress should simply borrow money from workers’ personal accounts to pay all promised Social Security benefits (going outside the accounts to credit markets only if necessary) and in exchange for the funds borrowed issue the accounts new long-term, inflation-protected bonds (TIPS) with no restrictions on resale in the secondary bond market. To paraphrase Alexander Hamilton, that debt would be to us a “national blessing.” [A longer version of this essay is available at www.ipi.org.]