The Investment-Based Social Security system will become self-financing by 2038. [See the discussion below.] Therefore, beginning in 2038 the amount of the 12.4 percent payroll tax going into real investments can begin to rise above the initial three percentage point level. As shown in Figure IV, the percentage of wages flowing into real investments through SI Accounts can rise over time, reaching a level of 8 percent around 2055. An additional 2 percent of payroll pays for disability insurance, and the payroll tax rate can be reduced over time from 12.4 percent to 10 percent. When fully phased in by 2055, the rate of real investment will be 8 percent of wages, yielding an expected basic benefit 2.5 times the amount promised by the current Social Security system for an average worker. The payroll tax will begin to decline after 2065 and reach the permanent level of 10 percent, thereby giving workers a 19 percent reduction in their Old Age Survivors and Disability Insurance (OASDI) taxes. By contrast, to maintain the solvency of the current Social Security system after 2065, we would be forced to raise the payroll tax by 48 percent or cut benefits by 33 percent.

In addition to retirement benefits, the Social Security payroll tax also funds family and survivors benefits and disability insurance. Those who elect to participate in the Investment-Based Social Security system are assured that these other benefits will be at least as great as under the current system.

Family and Survivors Benefits. Today, spouses (including some former spouses) and dependent children of retired workers receive a family benefit equal to 50 percent of the retired worker's expected Social Security benefit. At age 65, surviving spouses receive 100 percent of the deceased worker's expected Social Security benefit, with lower amounts paid to younger surviving spouses, dependent children and dependent parents. The total amount of benefits paid to all family members is capped at 188 percent of the retiree's base benefit, with additional amounts available for divorced spouses who were married to the retired worker for at least 10 years. Upon the worker's death, the surviving spouse or children receive a one-time, lump sum death benefit of $255. Also, family and survivor benefits are reduced dollar for dollar by the amount of Social Security benefits earned in the spouse's or child's own name.

Over a fairly short period of time, it seems probable that virtually every worker will opt into the investment-based system. Figure V shows the buildup of investment assets, net of retirement benefits paid, assuming all workers participate. The aggregate level of SI Accounts begins at zero in 2000 and grows to $1.6 trillion in just 10 years. The amount of assets builds up at an accelerating rate, reaching $10.5 trillion in 2040 and growing at a slower rate thereafter. The data on which Figure V and all other figures in this analysis are based are presented in Tables A-I and A-II in the appendix.

Figure VI shows the costs and benefits of transitioning from the existing debt-based system into an Investment-Based Social Security system. In the year 2000, the existing Social Security system will cost $394 billion. To maintain that level of funding and, at the same time, allow workers to invest 3 percent of wages in their SI Accounts will cost an additional $106 billion. The cost of the existing Social Security system is slated to rise to $699 billion in 2020. By that time, the investment-based system needs only an additional $92 billion to allow a real investment of 3 percent of wages and still pay the benefits of the existing system. The reason less additional funding is needed: roughly 20 percent of all benefits to new retirees in that year will be paid for by investments they made over the previous 20 years. By the year 2031, the investment-based system will cost roughly the same as the existing system. By 2040, 80 percent of the benefits paid to new retirees will be paid for by the investments they have made over the previous 40 years and the earnings on those investments. In that year, the cost of the existing system would have risen to $1,030 billion while the cost of the investment-based system will be only $920 billion.

By the year 2050, the cost of the existing Social Security system would be $1,150 billion and rising, whereas the cost of the new investment-based system would be back down to $861 billion and falling. In that year, the retirement benefits of the average new retiree would be fully funded through his or her SI Annuity. The remaining cost to taxpayers would consist of paying benefits to people who had retired previously, paying minimum benefits to those whose SI Annuities fall short of the guaranteed minimum and paying survivor and other benefits.

The initial transition costs of moving to an investment-based system are illustrated in Figure VI. The transition cost is 3 percent of wages in 2000 and declines to zero in 2031. In each of these transition years, the cost of the new system will exceed the cost of the existing system. But as worker investments build up under the new system and begin to fund retirement benefits, the transition cost falls off sharply. From the year 2031 on, the cost of the new system declines relative to the old system as real investments fund an ever-larger share of Social Security benefits. The cost of the existing system continues to rise because all benefits must be paid for with taxes on the wages of workers who are declining in number relative to the number of retirees. The benefits of moving to the new investment-based system beyond the year 2031 are also measured by the area between the two lines. Over time, the benefits of moving to an investment-based system greatly exceed the costs of the transition.

The Cash Flow Problem Associated with the Investment-Based System. From the previous analysis it is clear that, over any extended period, substantial savings accrue as a result of instituting an investment-based system and harnessing the power of compound interest to pay future benefits. In that sense, over a long period of time, there are no net transition costs in moving from the current debt-based system into an investment-based system. In fact, there are substantial savings. There is, however, a cash flow problem since the cost of moving to the investment-based system occurs in the first 31 years of implementation and the benefits flow in every year thereafter. Any serious proposal to institute an investment-based system for Social Security must contain a program to fund this cash flow problem.