[Economic Report of the President (2004)]
[Administration of George W. Bush]
[Online through the Government Printing Office, www.gpo.gov]





 
CHAPTER 6

Restoring Solvency to Social Security


Much of the Federal government's budget is dedicated to entitlement
programs, in which expenditures are determined not by discretionary
budget allocations but by the number of people who qualify. Reform
of entitlement programs remains the most pressing fiscal policy
issue confronting the Nation. With projected expenditures of $478
billion in 2003, Social Security is the largest entitlement program
and an appropriate place to begin.  Social Security is designed as a
pay-as-you-go system in which payroll taxes on the wages of current
workers finance the benefits being paid to current retirees. While
the program is running a small surplus at present, large deficits
loom in the future. Deficits are first projected to appear in 15
years; by 2080, the Social Security deficit is projected to exceed
2.3 percent of GDP.

The coming deficits in Social Security are driven by two
demographic shifts that have been in progress for several decades:
people are having fewer children and are living longer. The President
has called for new initiatives to modernize Social Security to
contain costs, expand choice, and make the program secure and
financially viable for future generations of Americans.

This chapter assesses the need to strengthen Social Security in
light of its long-term financial outlook. The key points in this
chapter are:
The most straightforward way to characterize the
financial imbalance in entitlement programs such as
Social Security is by considering their long-term
annual deficits. Even after the baby-boom generation's
effect is no longer felt, Social Security is projected
to incur annual deficits greater than 50 percent of
payroll tax revenues.
These deficits are so large that they require a
meaningful change to Social Security in future years.
Reform should include moderation of the growth of
benefits that are unfunded and can therefore be paid
only by assessing taxes in the future. A new system of
personal retirement accounts should be established to
help pay future benefits. The benefits promised to those
in or near retirement should be maintained in full.
The economic rationale for undertaking this reform in an
era of budget deficits is as compelling as it was in an
era of budget surpluses.

The Rationale for Social Security

All developed countries and most developing countries have publicly
administered programs to provide benefits for the elderly, including
programs to support surviving spouses and the disabled. Government
involvement in markets for goods or services is typically predicated
on a failure of private markets to achieve an efficient or equitable
result. There are three main problems in the market for providing
support to the elderly that justify a government role in old-age
entitlement programs.

First, a strictly private market to support old age would require
all individuals to choose the level of consumption that they would
like in retirement and to save accordingly. Some individuals may not
be capable of making the relevant calculations themselves and may not
be able to enlist the service of a financial professional to advise
them. For these people, Social Security provides a minimal level of
financial planning. Social Security requires people who otherwise
would not save for retirement to participate in a system that makes
them pay for insurance against old-age poverty. It also provides a
mechanism for everyone to share in the burden of taking care of those
who are truly in need of assistance.

Second, well-being in retirement is subject to two types of risk
that are not easily insured in private markets. The first risk is low
income during working years, which can lead to poverty in old age.
Low income may be caused by a specific event like disability, and
Social Security provides workers with disability insurance. Private
disability insurance plans exist but participation is quite low. Low
income may also be caused by other events beyond an individual's
control. However, these events do not lend themselves to private
insurance contracts because income can also be low for a variety of
reasons that are under an individual's control but that are difficult
for an insurer to observe (such as low work effort). Social Security
partially overcomes this problem through its progressive benefit
formula-- retirees with lower earnings during their working years get
benefits that are higher as a share of preretirement earnings.

The other risk to well-being in old age is the possibility that
retirees will live an unusually long time and thereby exhaust their
personal savings. To protect against this risk, a portion of the
retirement wealth that a worker has accumulated must be converted to
an annuity, a contract that makes scheduled payments to the
individual and his or her dependents for the remainder of their
lifetimes. The annuity payments should be indexed to inflation, so
that their purchasing power is not eroded over time. Inflation-indexed
annuities are a fairly new financial product, and even today,
relatively few people participate in the indexed annuity market. A
public system of Social Security, in which the government pays
benefits in the form of an annuity that keeps pace with the cost of
living, can help protect retirees from outliving their means to
support themselves.

Third, in other contexts, the government's fiscal policies are
designed to redistribute resources from high- to low-income
individuals. In most cases, such as the progressive income tax
schedule, income is defined based on an annual measure. Social
Security is unusual because it can redistribute income based on a
lifetime average of earnings. By doing so, Social Security more
accurately targets these transfers to the people who most need the
assistance. Individuals with higher lifetime average earnings receive
benefits from Social Security that are higher in dollar terms, but
lower as a percentage of their earnings, than do those with lower
lifetime average earnings.

All of these rationales are legitimate. Whether the U.S. system
actually meets these goals, and whether it does so in an efficient
and equitable manner, however, should be a subject of continued debate.
An essential part of this debate is that none of these rationales
require that Social Security be operated on a pay-as-you-go basis.
Long-term solvency can be restored by advance funding of future
obligations through personal retirement accounts. Personal retirement
account proposals can be and often have been designed to allow for
greater protection for surviving spouses and other vulnerable
groups. The President has taken an important step in this debate by
making the modernization and long-term solvency of Social Security
a prominent feature of his Administration's domestic policy agenda.

Understanding the Financial Crisis

In a pay-as-you-go system like Social Security, the benefits paid
to current beneficiaries are financed largely by the payroll taxes
collected from current workers. In any given year, the system will be
in balance when the income rate equals the cost rate. The income rate
is the total amount of tax revenue collected (from both the payroll
tax and the income taxation of Social Security benefits for moderate-
and high-income beneficiaries) divided by the total amount of payroll
on which taxes are levied. The cost rate is the total amount of
scheduled benefits divided by the total payroll on which taxes are
levied. The annual balance is the difference between the income rate
and the cost rate in a given year.

The impending financial crisis in Social Security is due to the
rapid growth in the cost rate relative to the income rate in the
future. This growth is attributable to two demographic factors that
have become critically important over the last half century: people
are having fewer children and living longer in old age. As a result of
these lower rates of both fertility and mortality, the size of the
elderly cohort will expand relative to the younger cohort over time.



Chart 6-1 compares the Social Security cost rate with the
dependency ratio, which is the number of beneficiaries per hundred
workers. The projections are based on the intermediate assumptions
made by the Social Security Trustees in their 2003 report. The
dependency ratio rises from 30.4 in 2003 to 55.2 in 2080, an increase
of 82 percent. Stated another way, the number of workers paying
payroll taxes to support the payments to each beneficiary will fall
from 3.3 workers per beneficiary in 2003 to 1.8 in 2080. With fewer
workers to support each retiree, it is not surprising that the cost
rate is projected to increase, in this case from 10.89 percent of
payroll in 2003 to 20.09 percent in 2080. This 84 percent increase is
almost identical to the rise in the dependency ratio. While changes
in productivity, immigration, interest rates, and other factors also
affect the long-term solvency of the program, changes in population
structure are at the center of the looming crisis.

Chart 6-2 graphs the single-year projections of Social Security's
income and cost rates, along with its annual balances. The solid curve
that rises over the period represents the cost rate (this is the same
curve as in Chart 6-1). The dashed line is the projected income rate,
which reflects revenue received by the Social Security trust funds
from the payroll tax of 12.40 percent plus a portion of the income
tax on current benefits. Income taxation on benefits currently being
paid generates an amount equal to 0.30 percent of taxable payroll.
Thus,




the income rate in 2003 was 12.70 percent. Because the income
thresholds at which Social Security benefits become taxable are not
indexed for inflation, a greater share of benefits become taxable over
time as the price level rises. In 2080, income taxation of benefits
is projected to generate 1.03 percent of taxable payroll, resulting
in an income rate of 13.43 percent.

The annual balance, the difference between the income rate and the
cost rate, is projected to deteriorate. For 2003, the annual balance
is 1.81 percent of taxable payroll (12.70  -  10.89). The annual
balance is graphed at the bottom of Chart 6-2 as a solid curve that
declines over the period. The substantial increase in the cost rate
relative to the income rate in the future causes this annual balance
to change from surplus to deficit by 2018 and to widen considerably
thereafter. In 2080, the annual balance will be
-6.67 percent of taxable payroll (13.43  -  20.09, as reported in the
Trustees Report, with the small discrepancy due to rounding).

Unless the Social Security system is reformed before that time,
the payroll tax would have to rise from 12.40 percent to 19.07 percent
to pay all benefits scheduled by current law, even with the
assumption that benefit taxation continues under current law to
provide a rising share of program revenues. Such an increase
represents an expansion of the payroll taxes associated with the
program of over 50 percent (Box 6-1).

The annual deficit of 6.67 percent of payroll is the most
straightforward way to represent the long-term fiscal challenge
confronting the Social Security program. To describe a proposed
reform as having restored solvency to Social Security, the reform
must greatly reduce or eliminate these annual deficits. The only
desirable way to restore solvency is to do so without

-----------------------------------------------------------------------
Box 6-1: The Retirement of the Baby-Boom Generation

It is common in public discussions to associate the financial
crisis in Social Security with the approaching retirement of the
baby-boom generation, those born between the years 1946 and 1964.
This explanation, however, is only partly correct. The problems
confronting Social Security are more fundamental than the aging
of an unusually large birth cohort. In 2080, for example, the
youngest baby boomer will be 116 years old, and almost all benefits
in that year will be paid to retirees who were born after the
baby-boom generation. Even with virtually no baby boomers among
the beneficiaries, Social Security in 2080 is projected to have
an annual deficit equal to 6.67 percent of its payroll tax base.

The retirement of the baby-boom generation does have an important
impact on the system's finances, as can be seen in Chart 6-1. The
period of rapid increase in both the dependency ratio and the cost
rate occurs during the two decades starting roughly in 2008 when
the baby-boom generation becomes eligible for retirement benefits.
Chart 6-2 shows that over this same period, the annual balance in
Social Security will deteriorate by over 5 percentage points of
payroll. If the retirement of the baby-boom generation were the only
source of Social Security's financial crisis, then the cost rate would
begin to decline as that generation passed away and the dependency
ratio fell.

As shown in Chart 6-1, however, the cost rate continues to climb
even as the baby boomers age and pass away. The dramatic increase in
the cost rate associated with the retirement of the baby-boom
generation is, in fact, a permanent transition to an economy in which
a higher ratio of beneficiaries to workers makes pay-as-you-go
entitlement programs more expensive to maintain. This transition would
be more apparent already were it not for the presence of the baby-boom
generation in the workforce today. The huge numbers of baby boomers in
the workforce have held down the ratio of beneficiaries to workers
over the past several decades. Judged from this point forward, the
retirement of the baby-boom generation does not cause the financial
crisis; it simply makes the long-term problem in the pay-as-you-go
system appear sooner rather than later.
-----------------------------------------------------------------------

continued reliance on general revenues. While these numbers are only
estimates and are revised over time, recent efforts by the actuaries
at Social Security to consider the uncertainty in the projections
show that there is essentially no chance that the system will be in
balance in the long-term (Box 6-2).

----------------------------------------------------------------------
Box 6-2: Long-Term Projections and Uncertainty

Recent experience with short-term forecasts has shown that there is
considerable uncertainty about how the economy will evolve. That
uncertainty is compounded over the 75-year period that the Social
Security actuaries must consider. Traditionally, the Trustees Report
has included projections based on three different sets of
assumptions--low cost, intermediate cost, and high cost. The low-cost
scenario has higher fertility rates, slower improvements in mortality,
faster real wage growth, and lower unemployment. All of these changes
work to reduce the projected deficits. The high-cost scenario changes
the assumptions in the opposite direction and results in larger
projected deficits.

Policy discussions seldom include any mention of the low- and
high-cost scenarios. Part of the reason is that these alternatives
are accompanied by no information on how likely they are to occur. In
the 2003 Trustees Report, a new method of dealing with uncertainty was
included in an appendix. The method, called stochastic simulation, is
based on the idea that each of the main variables underlying the
projection (like the interest rate or economic growth rate) will
fluctuate around the value assumed in the intermediate scenario. These
fluctuations are modeled by an equation that captures the relationship
between current and prior years' values of the variable and introduces
year-by-year random variation, as reflected in the historical period.
A stochastic simulation consists of many different combinations of
possible outcomes for the random variables. Each combination generates
a unique path for the key financial measures, each one analogous to
the single assumed path generated by the intermediate-cost scenario.
Taken together, these paths represent a wide range of possible
outcomes for Social Security.

Chart 6-3 shows the range of outcomes for the cost rate generated
by the simulation model. These simulations are based on the
assumptions and methods in the 2002 Trustees Report, when the
deficits reported in the last year of the projection period (2076)
were 1.11, 6.42, and 14.66 percent of taxable payroll in that year
for the low-, intermediate-, and high-cost scenarios, respectively.
Each curve, starting with the lowest, corresponds to a successively
higher percentile of the distribution of outcomes each year. In the
last year of the projection period, the median cost rate is 20.33
percent of taxable payroll, which is slightly higher than the value
of 19.84 percent based on the intermediate assumptions. Overall, 95
percent of the cost rates are between 14.53 and 28.98 percent of
payroll. Thus, the low-cost estimate of 14.24 and the high-cost
estimate of 28.51 correspond to very extreme outcomes in the overall
distribution.

Modeling the uncertainty underlying the demographic and economic
components of the projection is a large step forward in assessing the
future obligations of Social Security. The simulation model used in
the Trustees Report likely understates the variation that is possible
for future costs of Social Security. Nonetheless, the simulations show
that based on random year-to-year fluctuations, it is highly improbable
that the system will have a cost rate below its income rate in the
long-term. Uncertainty in the underlying projections only strengthens
the case for reform.

----------------------------------------------------------------------




Misunderstanding the Financial Crisis

Altough the Social Security program is operated on a largely
pay-as-you-go basis, discussions of the financial condition of the
program often focus on the trust funds out of which benefit payments
are made. There are two trust funds--one for the old-age and survivors
benefits and one for the disability benefits--that will be referred to
collectively as the ``Social Security trust fund.'' In a year when the
government collects more in payroll taxes than it needs to pay out in
Social Security benefits (net of the income taxes on benefits), surplus
revenues are allocated to the Social Security trust fund. The trust
fund is held in a portfolio that consists of special-issue Treasury
bonds.  The interest rate on the portfolio reflects the yields on
long-term Treasury bonds. In a year when Social Security benefit
payments exceed revenues, some of the bonds in the trust fund must
be redeemed to cover the gap.

In the 2003 Trustees Report, the trust fund ratio for the Social
Security program was reported as 288 percent for 2003. The trust fund
ratio is the proportion of a year's benefit payments that could be paid
with the funds available at the beginning of the year. Thus, a trust
fund ratio of 288 percent means that in 2003, the amount of bonds held
in the trust fund could have been redeemed to cover nearly three years
of Social Security benefit payments. A positive trust fund ratio is
the standard way of assessing the solvency of Social Security at a
point in time. A trust fund ratio of 100 percent is considered to be
an adequate reserve for unforeseen contingencies, such as an unexpected
drop in payroll tax collections in a particular year.

When the Trustees Report is released, the reaction in the popular
press almost always focuses on the date at which the trust fund is
projected to go to zero as an indicator of Social Security's financial
health. In the 2003 Trustees Report, this date was 2042, and this was
widely reported as good news because the prior year's report had
projected that date at 2041. The additional year before all of the
bonds are redeemed reflects higher annual balances in Social Security
through 2042 than were projected in the prior year's report.

Focusing on the date of trust fund exhaustion is inadequate as a
measure of Social Security's financial health because this date by
itself gives no indication of how dire the fiscal situation becomes
after the trust fund hits zero. When the trust fund is projected to
be exhausted in 2042, for example, the gap between the income and cost
rates on the Social Security program is projected to be 4.54 percent
of taxable payroll (or 37 percent of the revenues collected by the
payroll tax). If such a gap existed in 2003, it would be nearly $200
billion. Reform proposals that are based on pushing back the date when
the trust fund is exhausted by a few years will be insufficient to
address Social Security's long-term financial imbalance.

As a means of providing a longer-range summary of the finances of
the program, the Trustees Report also projects the 75-year actuarial
deficit in Social Security. Long-range actuarial projections are made
over 75 years because this is approximately the remaining lifetime of
the youngest current Social Security participants. The 75-year
actuarial deficit is equal to the percentage of taxable payroll that
could be added to the income rate for each of the next 75 years, or
subtracted from the cost rate for each year, to leave the trust fund
ratio at 100 percent at the end of the 75-year period.

In the Trustees Report for 2003, this 75-year actuarial deficit was
1.92 percent of taxable payroll using the intermediate assumptions, up
from 1.87 percent in the prior year's report. That is, in order to
have one year's worth of benefits left in the trust fund in 2077 (the
last year of the 75-year projection period starting in 2003), Social
Security payroll taxes would have to be 14.32 percent each year for
75 years.

The 75-year actuarial deficit is a widely used measure of the
system's financial condition. However, even this measure understates
the long-term challenge facing Social Security's finances. Although
an increase in the income rate of 1.92 percentage points in each of
the next 75 years leaves the trust fund with a positive balance at
the end of the 75-year period, the trust fund will rapidly decline to
zero in the years after 2077. This occurs because the payroll tax
increase of 1.92 percent does not cover the annual deficits of over
6.5 percent that are projected for those years.

Relying on the 75-year actuarial deficit as a guide to solvency is
only marginally better than considering the date of trust fund
exhaustion. A reform that purported to close the 75-year actuarial
deficit would be sufficient only to push the date of trust fund
exhaustion to a year just beyond the projection period.

The actuarial deficit over any finite period, even one as long as
75 years, can dramatically understate the financial imbalance in Social
Security when the program's annual deficits are getting wider over that
period. For example, the 2003 Trustees Report estimates that the
present value of the unfunded obligations for the program over the
next 75 years is $3.5 trillion. In other words, if this amount of money
were available today and invested at the rate of return that is
credited to trust fund assets, it would provide just enough to cover
the program's deficits over the next 75 years. However, the Trustees
Report also estimates that the present value of the program's unfunded
obligations over the infinite horizon--the next 75 years and all years
thereafter--is $10.5 trillion. The $7.0 trillion difference reflects
the continued annual deficits that persist after the first 75 years.
Thus, the first 75-year period represents only one-third of the
present value of the total shortfall.

A projection period limited to 75 years also biases the discussion
of potential reforms in favor of those that are based on
pay-as-you-go, rather than advanced, funding. Some reform proposals
would allow a portion of the payroll tax to be used to establish
voluntary personal retirement accounts (PRAs). People who establish
their own personal retirement accounts would be able to direct some
of their payroll taxes into their PRAs in exchange for accepting
lower benefits from the pay-as-you-go system in retirement. The
additional funding requirements to maintain benefits for current
retirees while allowing some of the payroll tax to be used for
personal retirement accounts for current workers necessarily appear
in the first 75 years. However, much of the benefit of advanced
funding--in terms of reduced obligations of the pay-as-you-go
system--occurs outside of the 75-year projection period.

Recognizing that even a 75-year actuarial deficit cannot fully
reflect the long-term financial shortfalls in Social Security, the
Trustees have increased their focus on the annual balance in the last
year of the 75-year projection period as a guide to the financial
shortfalls in the program. If the trust fund ratio is to continue
to play a role in discussions of solvency, then, at the very
least, the standard for restoring solvency to the program should
be to have not only a positive trust fund in the terminal year
of the projection period, but also a trust fund that is not declining
toward zero in that year.

The Nature of a Prefunded Solution

To restore solvency to Social Security on an ongoing basis, the
income and cost rates cannot be moving apart over time. If the income
and cost rates are moving together at the same level, then there is
no need for a large trust fund, because the program's annual balance
will be roughly zero in each year. As noted above, the annual deficit
is currently projected to grow to 6.67 percent of taxable payroll by
2080. Only by reducing annual benefits or increasing the payroll tax
(or the income tax on benefits) by a total of 6.67 percent of taxable
payroll can solvency be restored in the long term on a pay-as-you-go
basis.

If these benefit cuts or tax increases are not desired, then an
alternative is to allow the gap between the cost and income rates to
persist (provided that it is not increasing over time) and rely on
the investment income from a portfolio of assets to cover the gap.
Such a portfolio would have to be accumulated in the intervening
years, in order to prefund the difference between the program's
scheduled obligations and revenues.

In 1983, the last time a major reform of Social Security was
undertaken, the program was changed to begin accumulating annual
surpluses in the Social Security trust fund. In 2003, the trust fund
balance was $1.5 trillion. However, the intervening two decades
provide little assurance that the Social Security surpluses during
that time have increased the resources available to the government as
a whole to pay future benefits.

The balance in the Social Security trust fund has a clear meaning
as an accounting device. At any point in time, the trust fund balance
shows the cumulative amount of additional revenue--plus interest--the
Social Security program has made available to the Federal government
to spend on other purchases. The special-issue Treasury bonds in the
trust fund are IOUs from the rest of the government to the Social
Security program to cover its deficits in future years. The trust fund
balance shows the extent of the legal authority for the Social
Security program to redeem those IOUs in the future.
Administratively, the Social Security program is not authorized to
pay benefits unless the trust fund ratio is positive; that is, it can
only pay benefits to the extent that it has been a net creditor to
the rest of the government.

The question of what the government has done with the revenues made
available by past Social Security surpluses has important implications
for what the trust fund represents in economic terms and for the
design of Social Security reform. There are two competing conjectures
about the government's actions. The first is that the surpluses in the
Social Security program have had no effect on the surpluses or
deficits in the rest of the government's budget. If this is true,
every dollar that the government received from past Social Security
surpluses and thus allocated to the trust fund served to reduce the
amount of Treasury bonds held by the public by a dollar. In the
future, drawing down the trust fund when Social Security is projected
to run annual deficits simply involves selling the debt back to the
public so that when the trust fund is exhausted, the amount of debt
held by the public in the future will be the same as it would have
been had there not been any Social Security surpluses. Under this
conjecture about government budget policy, the Social Security
surpluses have been a source of higher national saving and the trust
fund represents real resources available to pay future benefits.

The second conjecture is that the surpluses in the Social Security
program have encouraged the government to run smaller surpluses or
larger deficits in the rest of its budget. If this conjecture is
true, the Social Security surpluses have not been used to repurchase
existing Treasury bonds held by the public but instead have been used
to pay for government expenditures, such as defense, health care, or
education. Drawing down the trust fund in future years will involve
selling Treasury debt to the public, as in the first case. However,
unless future government spending is reduced, the debt held by the
public will be higher than it would have been in the absence of Social
Security surpluses by the time the trust fund is exhausted. Under this
conjecture about government budget policy, the Social Security
surpluses have not resulted in higher national saving, and the balance
in the trust fund does not represent additional real resources
available to pay future benefits.

Analysts have argued in favor of both of these conjectures.
Determining which one is correct requires making an assumption about
what the government would have done in the counterfactual case that
Social Security had not run annual surpluses. The unified budget
deficit (including Social Security) has been the focus of budget
discussions for almost all of the last two decades. This provides a
strong prima facie case that government expenditures outside of Social
Security were higher due to the presence of Social Security surpluses
during this period.

Allocating Social Security surpluses to special-issue Treasury
bonds in the trust fund provides no guarantee that future Social
Security obligations are prefunded. It would therefore not be
appropriate to simply accumulate government bonds in a trust fund as
a way to restore solvency. One way to overcome the vagueness in trust
fund accounting is to require that the prefunding occur by allocating
a portion of Social Security's annual revenues to the purchase of
private rather than government securities and to treat these purchases
as annual expenditures of the Federal government. Doing so would
break the link between Social Security surpluses and the issuance of
debt by the Federal government. This would allow the Social Security
program to accumulate a portfolio of financial claims on private
sources to pay for future obligations.

Some simple arithmetic shows the size of the portfolio of private
securities that would be required to close the entire long-term annual
deficit in this manner. Suppose that investments in a portfolio of
stocks and corporate bonds earn a 5.2 percent expected return, net of
inflation and administrative costs. To obtain an income flow of 6.67
percent of taxable payroll (the annual deficit in 2080) would require
a portfolio of assets equal to 6.67/5.2 = 128 percent of taxable
payroll. In 2080, taxable payroll is projected to be 34.7 percent of
GDP, so that the required stock of assets would be equal to 44.5
percent of GDP. If such a fund existed in 2003, when GDP was estimated
to be $10.9 trillion, the fund would have a value of $4.9 trillion.
This calculation assumes that either taxpayers or beneficiaries will
absorb the financial risk associated with investments in corporate
stocks and bonds. Repeating the same arithmetic using a 3 percent real
interest rate--the projected return on the Treasury bonds in the
Social Security trust fund--shows that the fund would have to be $8.4
trillion.

For portfolios of this magnitude, prefunding by investing in
private securities would require that individuals establish their own
personal retirement accounts. To put these figures in some perspective,
as of November 2003, the net assets of all mutual funds in the United
States were estimated to be $7.24 trillion. Thus, in order to cover
the annual deficit in 2080 through prefunding, a portfolio the size
of at least two-thirds and possibly more than 100 percent of all
mutual funds would have to be accumulated. A portfolio of this size
is simply too large to be administered centrally without political
interference and without disruption to the capital markets.

In light of these issues, a Social Security reform plan should
have two components. First, it should restrain the growth of future
pay-as-you-go benefits for those not currently in or near retirement
to bring the cost rate of the program in line with the income rate in
the long term. Second, it should establish personal retirement
accounts for each worker. The personal retirement accounts serve a
dual purpose. First, because the accounts can be located outside of
the government's budget, the accumulation of assets in these accounts
would not provide any impetus for higher government spending in the
non-Social Security part of the budget. Second, the personal
retirement accounts would provide a way for individual workers to
accumulate assets to offset the reduction in their total retirement
income that otherwise would occur due to the lower benefits in the
pay-as-you-go part of the system.

Can We Afford to Reform Entitlements?

While Social Security's long-term solvency has been an ongoing
concern for over 25 years, the report of the 1994-1996 Advisory
Council on Social Security prompted a new round of policy discussions
that included serious proposals to prefund future obligations with
private securities. These discussions were bolstered by the
appearance of surpluses in the Federal government's budget and budget
forecasts during the late 1990s. Shortly after the President took
office in 2001, a bipartisan commission on Social Security was
established. The commission's final report discusses three reform
options that would involve the use of personal retirement accounts to
prefund a portion of future benefits.

Some critics of personal retirement accounts have suggested that
Social Security reform requires surpluses in the unified budget
(including Social Security) or even the non-Social Security portion
of the budget to begin investing in the accounts while maintaining
pay-as-you-go benefits to current retirees. Since the budget surpluses
forecasted a few years ago have not materialized, critics argue that
adding personal retirement accounts to Social Security is impossible
or impractical. In reality, the need to add resources to the Social
Security system is no less pressing now that the surpluses have
disappeared; indeed, it may be even more so. The change in the budget
outlook makes reform neither less necessary nor less economically
feasible.

As an illustration, consider the recent President's Commission's
Model 2, under the assumption that all eligible workers will
voluntarily choose to establish a personal retirement account
(thereby maximizing the transition costs to be discussed below). This
plan has two main components. First, it slows the growth of benefits
from the pay-as-you-go system by indexing future benefits to prices
rather than to wages. Prices generally increase more slowly than wages.
Second, the plan allows workers to receive a tax cut now, if they
place the tax cut into a personal retirement account, in exchange for
specific reductions in the pay-as-you-go benefits they would receive
otherwise. When workers choose this option, private saving is
increased. Under the conjecture that Social Security surpluses are
saved rather than spent, government saving is reduced and national
saving is essentially unchanged. However, the long-term solvency of
the pay-as-you-go system is maintained, and government and national
saving increase to the extent that having resources go into personal
retirement accounts rather than the Social Security trust fund
prevents the government from using Social Security revenues to pay
for non-Social Security expenditures.

The economic rationale for undertaking this type of Social Security
reform does not depend on the current budget situation. This is
clearly true with respect to the first component of reform--restraining
the growth of future pay-as-you-go benefits to a level that is
commensurate with future payroll tax revenues. The value of pursuing
this objective does not depend in any important way on whether, due to
prior economic and budgetary events not related to the reform, future
generations will be paying interest on a large or small stock of
public debt. If anything, easing the payroll tax burden on future
generations is more important if they face a greater interest burden.
Relying on personal retirement accounts also remains necessary.
Compared to government saving, saving in personal retirement accounts
gives workers greater freedom to prepare for their own retirement.
Saving in personal retirement accounts also ensures that the
additional resources being accumulated for Social Security are not
available to be tapped for additional government spending.

Even if both components of reform are still necessary, though, are
they feasible? Chart 6-4 shows the plan's effect on the unified budget
deficit and total government debt held by the public assuming that the
first contributions to personal retirement accounts are made in 2004.
Even under the favorable conjecture that Social Security surpluses do
not facilitate higher government spending outside of Social Security,
the deficit initially increases, but then falls as the reform is fully
phased in. At its maximum, in 2022, the incremental deficit increase
is less than 1.6 percent of GDP. The higher deficits in turn lead to
a greater stock of debt in subsequent years, followed by repayment.
The maximum increment to the debt is 23.6 percent of
GDP, in 2036.

The hump-shaped pattern for the impact of reform on the deficit
reflects the combined effects of the two parts of the reform. Personal
retirement accounts widen the deficit by design--they refund payroll
tax revenues to workers in the near term while lowering benefit
payments from the pay-as-you-go system in later years. After 2048, an
incremental surplus emerges as the benefit reductions phased in through
price indexation begin to outweigh the net effect of the personal
retirement accounts on the deficit.

Is this temporary increase in government borrowing a problem? Not
from an economic perspective. The increased borrowing does not shift
any burden to future generations. The tax cuts given to today's workers
are paid for by reductions in the share of their future benefits that
must be paid from future tax dollars. Nor are current workers harmed.
They save this money in their own accounts, which can give them
retirement income just as surely as if the government were promising
it to them.

While the government's budget situation does not affect the economic
necessity and feasibility of Social Security reform, under some
assumptions about the political constraints on the budget process, the
political feasibility and desirability of reform may be shaped by the
overall budget picture.

Reforms will lead to larger unified budget deficits in the near
term but smaller deficits in the long term. The presence of a deficit
in the non-Social Security part of the budget may make it more
difficult to persuade lawmakers to reform Social Security, if the
transition costs of the reform cause the deficit to eclipse a previous
record. However, avoiding Social Security reform will not keep
deficits in check. If nothing is done to reform Social Security, under
current projections, the growth of Social Security,




Medicare, Medicaid, and the interest on the borrowing required to
finance their growth will lead to unified budget deficits that
surpass previous records as a share of GDP.

Chart 6-5 shows the projected costs and revenues in the unified
budget under the assumption that no reforms are made to Social
Security. The projections are based on the President's policies in
the fiscal year 2004 budget, modified to include relief from the
alternative minimum tax. The chart assumes that all scheduled Social
Security benefit payments are made, financed through additional debt
after the trust fund is exhausted. The stacked areas represent total
scheduled Federal spending as a share of GDP. Even with nonentitlement
spending fixed at 8.1 percent of GDP and excluding interest payments,
Federal spending surpasses 20 percent of GDP in 2025, 25 percent in
2050, and 30 percent in 2080. The solid line shows total revenue. The
budget deficit, which is the height of the areas above the black line,
grows sharply in upcoming decades.

The impact of Social Security reform on the baseline deficit is
shown in Chart 6-6, which graphs the evolution of the deficit under
two scenarios: the baseline from Chart 6-5 in which no reform is
implemented and a reform that includes all of Model 2, with 100 percent
participation. Recall from Chart 6-4 that this reform causes the budget
deficit to increase temporarily before falling to a lower share of GDP
as the reform is fully phased in.



With the reform, the unified budget deficit reaches 5 percent of
GDP in 2019. Without reform, this deficit is reached instead in 2023.
The benefits of the reform appear over time, making a positive impact
on the Federal budget after 2048.

Policy makers concerned about the unified deficit will have to
decide how they will restrain Federal spending over the upcoming
decades--they will have to confront this question even if nothing is
done to reform Social Security. The benefit of reforming Social
Security is that it alleviates, to some extent, the financial burden
that unreformed entitlement programs will place on future generations.

Conclusion

The Nation must act to avert a long-foreseen future crisis in the
financing of its old-age entitlement programs. The crisis results
mainly from the fundamental demographic shifts to lower birthrates
and longer lives rather than the impending retirement of the baby-boom
generation. However, the scope for enacting meaningful reform will
disappear as the baby-boom generation begins to retire and an ever
greater share of the population sees its current income arrive in the
form of a government check. The design of the Social Security program
has failed to keep pace with emerging demographic realities.
The benefits promised to those currently in or near retirement must
be honored, but a new course must be set to ensure that Social
Security is viable and available to Americans in the future.

To do nothing at this point to restrain the growth of entitlement
programs would bequeath to future generations an increasing tax on
their income to support Social Security. The only way to avoid such
an outcome without reducing the living standards of future retirees
is to save more today. Greater saving will increase the capital stock
and increase the productive capacity of the economy so that it can
support those higher payments. The combination of reducing the
projected cost of taxpayer-financed benefits and shifting the revenues
into personal retirement accounts provides the best mechanism for
achieving that result.