March 25, 2004 | Backgrounder on Social Security
Social Security is the best-loved American government program, but how it works and is financed is almost completely unknown. Most Americans have a vague idea that they pay taxes for their benefits and that their benefits are linked somehow to their earnings. Many also know that the program is in trouble and needs to be "fixed" sometime soon to deal with the retirement of the baby boomers. Beyond this, their knowledge of the facts is severely limited and often colored by rumors and stories.
Most politicians exploit this lack of knowledge and limit their statements on Social Security to platitudes and vague promises. To make matters worse, reformers tend either to be content with similar platitudes or to speak in such detail that few outside the policy world can understand what they are saying. The simple fact is that today's Social Security is extremely complex, and any reform plan that is more than fine words will be similarly complex.
This paper attempts to simplify the reform debate by comparing various plans (including the current system) side by side. Each of the six sections of this paper compares how the current system and the reform plans handle a specific subject. Only reform plans that have been scored by Social Security's Office of the Chief Actuary are included in this comparison, using numbers contained in the 2003 Report of the Social Security Trustees. The six corresponding tables contain general reviews of aspects of the current system and the reform plans, with more details in the footnotes.
While looking at just one or two sections of special interest may be tempting, this approach would probably be misleading. For the best effect, each section should be considered together with the other sections in order to form a complete picture of the plan. Using simply one section by itself to judge an entire plan will not yield an accurate result.
Information in this side-by-side comparison is based on Social Security's scoring memos for each plan and conclusions that can be drawn from information contained in those memos. While there are many good points in the reform plans examined in this analysis, this is not an endorsement of any proposal by the author or The Heritage Foundation. Instead, this comparison provides details of specific plans. However, it would be wise for reformers to follow a set of general principles to ensure that any Social Security reform both resolves Social Security's problems and provides workers with greater retirement security. Those principles are listed below.
Social Security currently pays an inflation-indexed monthly retirement and survivors' benefit, based on a worker's highest 35 years of earnings. Past earnings are indexed for average wage growth in the economy before calculating the benefit. The benefit formula is progressive, meaning that lower-income workers receive a benefit equal to a higher proportion of their average income than upper-income workers receive. The program is expected to continue to collect more in payroll taxes than it pays out in benefits until about 2018.
Unused payroll taxes are borrowed by the federal government and replaced by special-issue Treasury bonds. After the system begins to pay out more than it receives, the federal government will cover the resulting cash flow deficits by repaying the special-issue Treasury bonds out of general revenues. When the bonds run out in about 2042, Social Security benefits will automatically be reduced to a level equal to incoming revenue. This is projected to require a 27 percent reduction in 2042, with greater reductions after that.
Representative Jim DeMint (R-SC) has introduced a voluntary personal retirement account (PRA) plan that would establish progressively funded voluntary individual accounts for workers under age 55 on January 1, 2005. The amount that goes into each worker's account would vary according to income, with lower-income workers able to save a higher percentage. For average-income workers, the account would equal about 5.1 percent of income.
The government would pay the difference between the monthly benefit that can be financed from an annuity paid for by using all or some of the PRA and the amount that the current system promises. The sum of the annuity and the government-paid portion of Social Security would be guaranteed at least to equal benefits promised under the current system, and 35 percent of PRA assets would be invested in government bonds to help pay for any Social Security cash flow deficits. This proportion would be reduced gradually in the future. General revenue money would be used to pay for additional cash flow deficits.
Senator Lindsay Graham (R-SC) has proposed a plan that would give workers under age 55 (in 2004) three options. (Workers above the age of 55 would be required to remain in the current system and would receive full benefits.)
Under Option 1, workers would establish PRAs funded with part of their existing payroll taxes, equal to 4 percent of pay up to a maximum of $1,300 per year. Workers' benefits would be reduced by changing the benefit indexing formula from the current wage growth index to one based on consumer prices. Over time, this change would reduce benefits for workers at all income levels, but the effect on lower-income workers would be eased by a mandated minimum benefit of at least 120 percent of the poverty level for workers with a 35-year work history. The government-paid monthly benefit would be further reduced to reflect the value of the PRA. This reduction would be calculated using the average earnings of government bonds so that, if the PRA earned more than government bonds, the total monthly benefit would be higher. Option 1 also raises survivor benefits to 75 percent of the couple's benefit for many survivors.
Option 2 is essentially the same as Option 1, but without PRAs. The government would pay all benefits for workers who choose this option. Option 2 includes both the basic benefit reduction and the minimum benefit requirement.
Option 3 pays the same level of benefits promised under current law, but workers who select this option would pay higher payroll taxes in return. Initially, the payroll tax rate for retirement and survivors' benefits would increase from 12.4 percent of income to 14.4 percent of income (counting both the worker's and the employer's shares of the tax). In subsequent years, the tax rate would continue to climb in 0.25 percent increments.
Representative Nick Smith (R-MI) has proposed a voluntary PRA plan that would create personal retirement savings accounts funded with an amount equal to 2.5 percent of income, paid out of existing payroll taxes. This would increase to 2.75 percent of income in 2025 and could become larger after 2038 if Social Security has surplus cash flows. Retirement and survivors' benefits would be reduced by an amount equal to the value of lifetime account contributions plus a specified interest rate.
The Smith plan would also make many changes in Social Security's benefit formula, mainly affecting middle-income and upper-income workers. These changes would eventually result in most workers receiving a flat monthly benefit of about $550 in 2004 dollars. It would also gradually increase the retirement age for full benefits and require that all newly hired local and state workers be covered by Social Security. The Smith plan transfers $866 billion from general revenues to Social Security between 2007 and 2013 to help cover cash flow deficits and allows additional general revenue transfers when needed after that.
Peter Ferrara, Director of the International Center for Law and Economics, has proposed a plan that would create voluntary PRAs that would be funded according to a progressive formula that allows lower-income workers to save a higher proportion of their payroll taxes than upper-income workers. Average-income workers could save about 6.4 percent of their income. Workers would be guaranteed that the total of their PRA-generated benefits and government-paid monthly benefits would at least equal the benefits promised under the current system.
Any Social Security cash flow deficits that remain would be financed through general revenue transfers equal to a 1 percent reduction in the growth rate of all government spending for eight years, the corporate income taxes deemed to result from the investment of personal account contributions, and issuing about $1.4 trillion in "off-budget" bonds. Under the Ferrara plan, these bonds would be considered a replacement for the existing system's unfunded liability and thus would not increase the federal debt.
Peter Orszag, Senior Fellow at the Brookings Institution, and Peter Diamond, Institute Professor of Economics at the Massachusetts Institute of Technology, have developed a plan that does not include any form of PRA or government investment of Social Security trust fund money in private markets. Instead, it gradually changes the benefit formula to reduce benefits for moderate-income and upper-income workers and requires that all state and local government workers come under Social Security. It would also gradually reduce benefits by raising the age at which workers could receive full benefits. Workers could still retire earlier, but at lower benefits. Benefits would increase for lower-income workers, widows, and the disabled.
In addition, the plan would gradually increase the payroll tax for all workers from the current 12.4 percent of income to 15.36 percent of income in 2078. It would also raise the earnings threshold on Social Security taxes--thus requiring higher-income workers to pay additional payroll taxes--and impose a new 3 percent tax on income above the earnings threshold. Workers would not receive any credit toward benefits for income covered by this new tax.
What Is This,
and Why Is It Important?
Allowing workers to invest a portion of their Social Security taxes is the only alternative to raising Social Security taxes or reducing Social Security benefits. However, personal retirement accounts are not all equal. The money that goes into the PRAs could come from diverting a portion of existing Social Security taxes or from some other source. (See Table 1.)
Similarly, the size of the accounts (usually expressed as a percentage of the worker's pay) is important. While larger accounts would temporarily increase the amount of additional funds required to pay benefits to retirees, they would also accumulate a pool of money faster than smaller accounts and finance a greater portion of benefits in future years. This can reduce the amount of additional tax dollars needed in future decades.
Finally, how the PRAs are invested is important. Even though they show steady growth over time, stocks and commercial bonds are generally more volatile than government bonds. Investing a portion of the PRAs in government bonds makes the accounts slightly less volatile while providing some of the additional dollars needed to pay benefits to current retirees.
What Is This,
and Why Is It Important?
Other than creating personal retirement accounts that allow workers to self-fund all or a portion of their Social Security retirement benefits, most reform plans deal with the program's coming deficits by either changing the level of retirement benefits promised or finding ways to increase program revenues. This section examines how various reform plans treat promised retirement benefits. (See Table 2.)
Social Security uses a complex formula to calculate an individual worker's retirement benefits. Subtle changes in this formula can cause a large change in benefits over time. For instance, changing how past income is indexed to a constant purchasing power will have only a minor impact for the first several years. However, the effect is cumulative and after several decades will result in major changes in benefits.
Similarly, seemingly minor changes in "bend points"39 or other aspects of the benefit formula can, over the long term, cause major changes in benefits for upper-income and/or moderate-income workers. It is even possible to use the benefit formula to approximate an increase in the full retirement age without actually raising it. Thus, a plan could still allow workers to qualify for "full retirement benefits" at 65, 66, or 67 but award them full retirement benefits (as defined under the current system) only if they wait to retire until a later age.
The first question that any plan must answer is whether it would pay the full level of benefits promised under the current system. If so, it must deal with how to pay the cost, since the current system cannot afford to pay for all of the promised benefits. Other important questions include whether the plan proposes benefit changes (usually reductions) if workers do not choose to have a personal retirement account, protects lower-income workers (who more often have an interrupted work history) by instituting some sort of minimum benefit level, and/or addresses the low benefits for certain lower-income, widowed, and disabled workers under the current system.
What Is This,
and Why Is It Important?
Increasing Social Security payroll taxes would be one way to pay projected cash flow deficits. This method is closer to the self-funding that has characterized the system so far, but raising payroll taxes has significant drawbacks. Alternatives to payroll tax increases include instituting some form of personal retirement account to increase the return on taxes, reducing benefits, and using significant amounts of general revenue money to cover Social Security's cash flow deficits. (See Table 3.)
Currently, all workers pay 5.3 percent of
their income to pay for Social Security retirement and survivors
benefits. In 2004, this tax will be paid on the first $87,700 of an
Employers match this tax for a total of 10.6 percent of each
worker's income. In addition, both employer and employee pay an
additional 0.9 percent of the worker's income (1.8 percent total)
for Social Security disability benefits. Thus, the employer and
employee pay a total Social Security payroll tax of 12.4 percent.41
Additional payroll taxes could be collected in three ways:
What Is This,
and Why Is It Important?
Both the current Social Security system and every plan to reform it will require significant amounts of general revenue money in addition to the amount collected through payroll taxes. This additional money is necessary to reduce the difference between what Social Security currently owes and what it will be able to pay.
In the reform plans, the transition cost represents a major reduction from the unfunded liability of the current program. Even though the reform plans are expensive, all of them would require less additional money than the current system. However, both the amount and the timing of this additional money would vary depending on the plan. (See Table 4.)
Present value reflects the idea that a dollar today has more value to a person than that same dollar has sometime in the future. It gives an idea of when the additional money is needed by giving greater weight to money needed in the near future than to an equal amount needed further in the future. In addition to showing the amount of money needed, a higher present value number indicates that money is needed sooner rather than later.
The sum of the deficits indicates the total amount of additional money that will be needed. This measure gives $100 needed today the same weight as $100 needed in 15 years. This measure adds up only the future cash flow deficits; it does not include cash flow surpluses because the government does not have any way to save or invest that money for future use. Using both of these measurements gives a better picture of the situation than using just one.
Paying for the current system or any of the reform plans will require Congress to balance Social Security's needs against those of the rest of the economy. In general, as more additional dollars are needed for the current system or a reform plan, less money will be available for other government programs and the private sector.
As this burden on the general federal budget increases and persists, Congress would find it increasingly more difficult to come up with that money, and it would become increasingly less likely that such a plan would really be paid for on schedule. This is especially true for the current system, which will incur the massive deficits to pay all of the promised benefits.
The numbers used in Table 4 were calculated by the Office of the Chief Actuary using static scoringmethods. Dynamic scoring would give a more complete picture of the economic effects of each plan, allowing analysts to compare a plan's ability to create jobs, increase savings, and generate economic growth. In many cases, economic growth associated with a reform plan could increase or reduce the amount of general revenue required to finance it. Regrettably, the Social Security Administration does not offer dynamic scoring at this time.
What Is This,
and Why Is It Important?
Both the current Social Security program and all of the proposed reform plans will require large amounts of general revenue money to cover the annual cash flow deficits. Exactly when that money is first needed, how many years it will be needed, and the total amount that will be needed varies from plan to plan. Avoiding use of general revenue money would require either reducing Social Security benefits enough to eliminate the annual deficits or imposing new taxes to generate sufficient revenue. Neither the current system nor any of the proposed reform plans comes close to closing the gap.
plans do specify sources for the needed general revenues (See Table
5.), but these are handicapped by the fact that no Congress can
bind the hands of a future Congress. Thus, even if Congress did
pass a plan that specified the source of the needed general
revenues, a future Congress could change the plan by a majority
vote. The only way to avoid this uncertainty would be for Congress
to pass and the states to ratify the plan as a constitutional
amendment--which would be prohibitively difficult.
In short, both the current system and all known reform plans would have to find the necessary general revenues from some combination of four sources: borrowing additional money, collecting more taxes than needed to fund the rest of the government, reducing other government spending, or reducing Social Security benefits more than is called for under either current law or any of the reform plans.
The most important thing to remember is that the existing Social Security system and the reform plans all face this problem. This is not a weakness that is limited to PRA plans or any other reform plan. The only question is when the cash flow deficits begin and how large they will be.
Current law makes no provision for funding Social Security's unfunded liability. The program has no credit line with the U.S. Treasury, and when its trust fund promises are exhausted, current law will require it to reduce benefits.
While some press releases connected with Representative DeMint's plan suggest that some of its general revenue needs could be generated by reducing the growth of federal spending, no language specifying where the general revenues would come from is included in his legislation.
Senator Graham's plan in-cludes a commission that would recommend reductions in corporate welfare and redirect the savings to reduce his plan's unfunded liability. At best, a reduction in corporate welfare would generate only part of the needed general revenue. The commission would produce a legislative proposal that would then be considered by Congress.
Because the commission would be created by the same legislation that implements Graham's Social Security reforms, its recommendations could not even be considered until after the plan is enacted. As a result, passage of the Graham plan does not guarantee that these revenues would be available. Regardless of what the commission recommended, a future Congress could reject the proposed cuts in corporate welfare. In that case, Congress would have to come up with another method to raise the needed revenue.
First, it would mandate a 1 percent reduction in the growth of all federal spending (including entitlements such as Social Security) for at least 8 years and redirect that revenue to Social Security. Since Congress cannot legally force a subsequent Congress to follow a set course of action, the only enforcement mechanism available is a constitutional amendment. As a result, the Ferrara plan simply
appropriates to Social Security the amount of revenue that would result if Congress were to reduce spending growth. In practice, a future Congress could choose not to reduce spending growth and, instead, just let the deficit grow larger or generate the necessary revenue in some other way.
Second, the Ferrara plan would transfer to Social Security the amount of corporate income taxes that could potentially result from the investment of personal accounts in corporate stocks and bonds. This is not a new or higher tax. This transfer is intended to reflect the taxes that would be paid at the current 35 percent corporate tax rate. Since SSA does not conduct dynamic scoring, this transfer is based on the static assumption that two-thirds of the stocks and bonds held through personal accounts reflect domestic corporate investment.
Third, the Ferrara plan would borrow about $1.4 trillion in special off-budget bonds. However, there is no practical way to create off-budget bonds that would not count against the federal debt. Even if there were, such a move would reduce the amount of transparency in the federal budget.
While the Orszag-Diamond plan includes both some benefit reductions and benefit increases for widows, the disabled, and low-income workers, the two elements of the plan are roughly equal. It reduces Social Security's unfunded liability using tax increases contained in the plan, including an increase in the payroll tax rate, a gradual increase in the amount of income subject to Social Security taxes, and a new 3 percent tax on any salary income not subject to Social Security taxes.
What Is This,
and Why Is It Important?
In the long run, a reform plan should do more than just preserve the current Social Security system with its many flaws. While a key requirement of any reform plan is to provide a stable, guaranteed, and adequate level of benefits at an affordable cost, it should do more. (See Table 6.)
The current system fails to allow workers to build any form of nest egg for the future. Instead, it is the highest single tax for about 80 percent of workers. In return, each worker receives a life annuity that ends with the death(s) of the worker, the surviving spouse (if there is one), or young children (if any). In today's world, where two-earner families are increasingly the norm, the current system even limits survivor benefits to the higher of either the deceased spouse's benefits or the surviving spouse's benefits. Whichever account is lower, no matter how long that spouse worked, is marked paid in full and extinguished.
At a minimum, a reform plan should allow workers to pass on some of what they earned and paid in Social Security taxes to improve their spouse's retirement benefits. It should also allow workers the flexibility to use their entire account for retirement benefits or take a smaller retirement benefit and use the balance to pay for a grandchild's college education, start a small business, or pass on money to a later generation.
In judging whether each proposed reform would be better for America's workers, readers may differ sharply. However, while most summaries and studies examine Social Security reform from the viewpoint of federal budget impact, tax rates, and the survivability of the system, few consider the overall impact of reform on the workers it was designed to benefit in the first place. Social Security should not be reformed or "saved" for its own sake, but only if it more effectively provides the benefits workers need at a price they can afford.
David C. John is Research Fellow in Social Security and Financial Institutions in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.
1. Stephen C. Goss, Chief Actuary, Social Security Administration, "Estimated Financial Effects of H.R. 3177, the `Social Security Savings Act of 2003,'" memorandum to Representative Jim DeMint, September 26, 2003, at www.ssa.gov/OACT/solvency/DeMint_20030926.html (February 28, 2004).
2. Chris Chaplain, Actuary, and Alice H. Wade, Deputy Chief Actuary, Social Security Administration, "Estimated OASDI Financial Effects of `Social Security Solvency and Modernization Act of 2003' introduced by Senator Lindsey Graham," memorandum to Stephen C. Goss, November 18, 2003, at www.ssa.gov/OACT/solvency/LGraham_20031118 (February 28, 2004).
3. Chris Chaplain, Actuary, and Alice H. Wade, Deputy Chief Actuary, Social Security Administration, "Estimated Long-Range OASDI Financial Effect of a Proposal Developed by Representative Nick Smith," memorandum to Stephen C. Goss, September 10, 2003, at www.ssa.gov/OACT/solvency/NSmith_20030910.html (February 28, 2004).
4. Stephen C. Goss, Chief Actuary, Social Security Administration, "Estimated Financial Effects of `The Progressive Personal Account Plan,'" memorandum to Peter Ferrara, December 1, 2003, at www.ssa.gov/OACT/solvency/PFerrara_20031201.html (February 28, 2004).
5. Stephen C. Goss, Chief Actuary, Social Security Administration, "Estimates of Financial Effects for a Proposal to Restore Solvency to the Social Security Program," memorandum to Peter Diamond and Peter Orszag, October 8, 2003, at www.ssa.gov/OACT/solvency/DiamondOrszag_20031008.html (February 28, 2004).
6. Alternative sources of funds for the PRAs would require either additional savings or additional taxes. In both cases, the result would lower the overall rate of return because the worker would be paying more.
7. Some plans offer additional investment options. All stock investment options would be invested in stock index funds, not individually selected stocks. Similarly, government and commercial bond investments would be in broadly based pools of bonds rather than in individually selected bonds.
8. At least initially, all accounts are managed by a single entity offering limited investment choices. This structure significantly reduces administrative costs. In all cases, the fund management would be contracted out to a private manager by a government supervisory agency, similar to how federal workers' Thrift Savings Plan is managed. Some plans would allow workers to select another funds manager when their accounts reach a certain size.
10. This is the average account size for median-income workers. Lower-income workers could put a higher percent of their income into a PRA than higher-income workers. This progressive feature approximates the current Social Security system, which gives lower-income workers higher benefits for their taxes than higher-income workers receive.
12. Workers could contribute up to an additional $2,000 and receive a partial tax credit. They could also roll over money from other retirement accounts into their Social Security PRA. Low-income workers could receive an additional $300 from the government to their account.
17. Some experts are concerned that PRAs could result in lower benefits for low-income workers because they may have intermittent work histories. A guaranteed minimum benefit would provide additional protections for these workers.
18. For many years, experts on all sides of the debate have criticized the current system for providing inadequate benefits to a surviving spouse. They have also expressed concern that benefit changes for retirees and their survivors would be unfair to those receiving benefits under Social Security's disability program.
19. Under current law, Social Security will automatically reduce the benefits when the trust fund runs out in 2042. From that point on, Social Security would only pay benefits equal to its cash flow. The SSA estimates that this would immediately reduce benefits by 27 percent in 2042, with larger reductions in subsequent years.
20. Under current law, Social Security uses the same benefit formula to compute both retirement and disability benefits. When a disabled worker reaches full retirement age, the worker's benefits switch from being paid by the disability program to being paid through the retirement and survivors program.
23. Under the DeMint plan, disability benefits would not changed. The PRA of a deceased spouse would be added to the surviving spouse's PRA account after subtracting the amount needed to pay benefits to any survivors. If there are no survivors, the PRA would go to the worker's estate.
24. The Graham plan allows workers to lock in the current level benefits if they are willing to pay higher payroll taxes. If a worker chooses that option, payroll taxes would climb to 14.4 percent from the current 12.4 percent (counting both the employer and employee shares). It would increase further if a higher rate were necessary to pay benefit levels promised under current law. If a worker chooses another option, the worker's benefits would be reduced below the level promised by the current system.
25. The Graham plan would change how benefits are calculated when the worker retires, from the current method of indexing past earnings to wage growth in the economy to indexing based on inflation and price growth. This would gradually reduce benefits. Workers who choose to pay higher payroll taxes would avoid this reduction.
27. Under the Graham plan, disability benefits would not change from the current level. However, when a disabled worker reaches retirement age, the retirement benefits could be reduced to reflect changes in benefit formulas during times when the worker was not on disability. Surviving spouses would receive at least 75 percent of the benefit received when both spouses were alive, subject to some restrictions. If a worker chooses to have a PRA, the balance in that account would be transferred to the spouse's account when the worker died.
29. The Smith plan would make many changes to Social Security's benefit formula, primarily affecting middle-income and upper-income workers. For most workers, it would eventually result in a flat monthly benefit of about $550 per month (in 2004 dollars). It would also gradually increase the retirement age for full benefits.
30. While there is no set minimum benefit as contained in some other plans, lower-income workers would receive additional credits into their PRA and any benefit reductions that would otherwise affect them would be limited by a maximum reduction.
31. Under the Smith plan, lower-income disabled workers would not be affected by reductions to the benefit formula. Disabled workers with incomes above the threshold would see a minimum benefit level. Benefits could be adjusted.
34. The Ferrara plan does not address disabled workers in any detail. However, because the plan is based on the combined trust fund of both the retired/survivors program and the disability program, one could assume that disability benefits would be paid out of a PRA and that current disability benefits would also be guaranteed. Spouses are guaranteed current law benefits. Upon the death of one spouse, the balance in the PRA would be transferred to the surviving spouse's PRA.
35. The Orszag-Diamond plan includes a number of benefit reductions that would especially affect upper-income and middle-income retirees. It also effectively increases the retirement age for all income levels by making it impossible for workers to get their full benefits unless they choose to delay retirement beyond the current normal retirement age.
36. The Orszag-Diamond plan would gradually change the benefit formula to reduce benefits for moderate and upper-income workers. Although workers could still retire earlier, they would receive lower benefits.
37. The benefit for minimum wage workers with at least 35 years of work history would be phased in over time, with the benefits reaching at least the poverty level by 2012. This benefit would gradually increase in subsequent years.
38. Surviving spouses would be guaranteed a benefit of at least 75 percent of the amount that the couple received when both were still living. Benefits for disabled workers would not be affected by changes to retirement and survivor's benefits.
39. The benefit formula used by the current Social Security system develops an "average indexed monthly earnings" for each worker by indexing his or her highest 35 years' earnings covered by Social Security taxes according to the growth in wages that has occurred between the date they were earned and the date that the benefit calculation is being made. In the next step, the actual retirement benefit is calculated. In 2003, the formula paid benefits equal to 90 percent of the first $606 of a worker's average indexed monthly earnings, 32 percent of the amount between $606 and $3,653, and 15 percent of any indexed earnings above $3,653. The divisions between the 90 percent, 32 percent, and 15 percent levels are called bend points.
41. Although the federal government considers the employer's matching share as a separate item, most employers add their portions of the Social Security tax to a worker's salary when calculating the true cost of an employee.
43. Benefits are currently calculated using all of the income upon which the worker paid Social Security taxes. Plans marked "yes" would break this link. Many analysts on both the left and the right believe that breaking this link would be the first step toward changing Social Security into a welfare system.
46. Between 2005 and 2063, the Orszag-Diamond plan would gradually raise the amount of income subject to Social Security taxes by 0.5 percent points annually, in addition any increases under the current formula.
47. The Orszag-Diamond plan would impose a new tax of 3 percent on all income above the current taxable earnings limit; however, workers would not receive any credit toward Social Security benefits for income subject to this tax. The plan's authors estimate that this tax would increase to 3.5 percent by 2080.
49. Present value is a measurement of the amount of money that if invested today would finance future benefit payments. For example, if Social Security will owe $1,000 in 30 years and assuming that an investment would earn an average of 3 percent every year after inflation (a growth rate equal to what government bonds pay), the present value of the $1,000 due in 30 years is $412. (Invested at a 3 percent interest rate, $412 will grow to $1,000 in 30 years.) Because the money is assumed to grow over time, a dollar needed in the future counts as less in present value terms than a dollar needed today.
50. This is another measurement of the amount of additional money needed. In this case, the "additional amount" is the total amount of additional federal revenues required to fund the transaction. Surpluses (if any) are not included since current law does not allow the government to do anything with excess cash except spend it. Under this method, if Social Security needed $1 billion in 2005 to make benefits payments and another $1 billion in 2025, the additional amount would be the total of $2 billion.
51. The current Social Security system will require additional revenues in excess of $100 billion annually for 57 years, in excess of $300 billion annually for 45 years, and in excess of $400 billion annually for 33 years.
52. The DeMint plan will require additional general revenues in excess of $100 billion annually for 35 years. It will not require additional general revenues in excess of $300 billion annually in any year.
53. The Graham plan will require additional general revenue in excess of $100 billion annually for 32 years. It will not require additional general revenues in excess of $300 billion annually in any year.
54. This scoring assumes that 100 percent of eligible workers would choose to have a personal retirement account. The plan was also scored for 67 percent and 0 percent participation, both of which require less additional general revenue dollars.
55. The Smith plan will require additional general revenues in excess of $100 billion annually for 24 years. It will not require additional general revenues in excess of $300 billion annually in any year
56. Supporters of the Ferrara plan believe that the maximum annual deficit would be $49 billion in 2005 if the various methods that it uses to finance the transition are taken into consideration. However, this section of the paper deals only with the size of the necessary general revenue transfers, not how the plans propose to finance them. The next section discusses how plans propose to find the necessary general revenues.
57. The Ferrara plan will require additional general revenues in excess of $100 billion annually for 47 years, in excess of $300 billion annually for 30 years, and in excess of $400 billion annually for 20 years.
58. The Orszag-Diamond plan will require additional general revenues in excess of $100 billion annually for 45 years. It will not require additional general revenues in excess of $300 billion annually in any year.
59. This is the most subjective measure in this paper. It compares the retirement benefits offered under the reform plan to the amount of general revenue money that is needed to finance it, while taking into consideration both changes in taxes (if any) and the ability to build a nest egg. It compares these to what the current system would be able to pay rather than what it promises.
61. Under the DeMint plan, workers may annuitize either all of their PRAs or 35 percent. However, the total income generated from a combination of the annuity and any government-paid benefit must total at least a poverty level income. If annuitizing 35 percent of the PRA does not reach this level when combined with a government-paid benefit, then enough of the PRA must be annuitized to reach that level. Any money not annuitized may be used for any purpose,
62. The DeMint plan would allow workers to improve their retirement benefits over the current system. While it has somewhat higher general revenue costs than other plans, it also guarantees a higher level of benefits and allows workers to build a nest egg for the future. The plan contains an innovative method to reduce general revenue costs by requiring workers with personal accounts to invest a certain proportion of their account in government bonds, and it contains the most progressive personal account contribution rate of any of the plans. This would allow it to closely match the current system's progressive benefit levels through its account structure alone.
63. The Graham plan requires workers to annuitize enough of the personal retirement account to provide a poverty-level income for their households, when combined with the government-paid monthly benefit. Any remaining amount may be used for any purpose.
64. The flexibility contained in the Graham plan offers workers much more control over their futures. Each worker would have the opportunity to decide the level of benefits, cost, and risk. The plan offers benefits above the current system at a relatively low general revenue cost. It also allows workers the opportunity to build a nest egg.
65. Under the Smith plan, the treatment of a deceased worker's PRA is less clear. It appears that upon a worker's death, any remaining PRA balance would be transferred to the PRAs of any survivors or the worker's estate if there are no survivors.
66. The Smith plan requires workers to annuitize enough of the personal retirement account to provide a poverty-level income for their households, when combined with the government-paid monthly benefit. However, unlike most other personal account plans, workers must either annuitize the full PRA or make regular withdrawals from amount remaining in the PRA after annuitization, although the worker may only draw enough so that the total lasts throughout retirement. The worker does not have the option of using remaining money for other purposes than retirement.
67. Overall, the Smith plan's benefit to workers is questionable. While general revenue costs are comparatively low, the eventual severe reduction in benefits will transform Social Security into a low flat-rate benefit that is substantially less than the current system's average benefit level. The plan also offers only very limited opportunities to build a nest egg. If the Smith plan were enacted, some form of almost mandatory occupational pension savings plan would be needed to provide workers--and especially lower-income workers--with a decent standard of living during retirement.
68. Under the Ferrara plan, workers must buy an annuity that pays a monthly amount equal to the worker's Social Security benefits under the current system. If the account is not large enough to buy the annuity, the federal government would make up the difference. Any money not annuitized may be used for any purpose.
69. The Ferrara plan could potentially improve workers' retirement benefits, but its substantial costs and financing mechanism may offset these benefits. While workers would be guaranteed at least the same level of retirement benefits as under the current system and would have the opportunity to build a nest egg, the higher general revenue taxes necessary to finance the plan make an overall improvement questionable.