"... prefunding Social Security and Medicare benefits would avoid the need for any tax increases 20 years or more from now."
Because of increasing life expectancy, if no changes are made to the current Social Security and Medicare systems, then based on estimates by the Congressional Budget Office, tax rates to fund promised benefits would have to more than double by the year 2050. But in The Economics of Prefunding Social Security and Medicare Benefits (NBER Working Paper No. 6055), NBER President Martin Feldstein and Faculty Research Fellow Andrew Samwick show that prefunding Social Security and Medicare benefits would avoid the need for any tax increases 20 years or more from now. Moreover, they show that the enhanced capital investment generated by prefunding would eventually increase the wealth of Americans by more than 30 percent.
Feldstein and Samwick analyze prefunding Social Security and Medicare with a system of mandatory savings accounts. They focus on the Social Security system but also apply their approach to Medicare. In the long run, the current pay-as-you-go (PAYGO) Social Security system would be replaced by a fully funded system in which employees over the age of 30, and their employers, would make annual payments into Personal Retirement Accounts (PRAs). Individuals, though required to make such payments, would choose the particular stocks and bonds in which the payments would be invested. The interest, dividends, and capital gains on such accounts would not be taxed at any time. Also, the government would contribute to each account the extra corporate income tax collected as a result of the increased saving. The accounts would earn a real return of about 9 percent, the rate that Feldstein and Samwick estimate as the pretax real return on investment in the U.S. economy. When the individual reaches retirement age, the accumulated fund would be used to buy an annuity that earns the same rate of return.
Because savings in the PRAs would earn such a high rate of return, the researchers estimate, benefits promised by the current Social Security program could be funded in the long run by employee-employer payments equal to just 2 percent of covered wages. The current payroll tax, in contrast, is 12.4 percent of covered wages, and a whopping 18.75-percent rate would be required in 75 years under the current system. The dramatic difference in the contribution rates reflects the power of compounding: funds growing at a real rate of 9 percent allow a given level of benefits to be funded with a much lower initial investment.
Feldstein and Samwick also analyze a 25-year transition to such a fully funded system. During this transition, the combined tax and PRA contribution rate would have to rise, starting at a peak of 14.4 percent in the first year, but would then decline to 10.71 percent by the 25th year, well below the current tax rate of 12.4 percent.
Although prefunding would produce large net benefits for American residents, some age groups would suffer minor losses. The biggest losses would be for those who are middle-aged at the start of the transition because they would be paying partly to prefund their own benefits and partly to fund benefits for the elderly. Those aged 40 when the transition begins would lose the most, but only about $5,000 over their whole lifetime. Combining their losses with their children's gain gives a much different picture, though. A husband and wife aged 40 with two children below age 20 would have a large net gain. That is because their funds would be earning a high real return rather than the low rate of return that the current Social Security system yields.
Feldstein and Samwick conclude: "Designing an appropriate way to finance the retirement and health care benefits of the aging population is probably the most important challenge to government finance in the decades ahead. If it is done wisely, the aged will have comfortable retirements and the advantages of improving medical technology while the working population will avoid the explosive growth of taxes that could otherwise occur."