Full text of DOI:10.1002/(SICI)1520-6688(200021)19:2<263::AID-PAM5>3.0.CO;2-H

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 263
Risky Business?
Evaluating Market Risk of
Equity Investment Proposals
to Reform Social Security
Christian E. Weller
Abstract
A number of options have been proposed to address the expected financing shortfall
of Social Security in the next century. Most basic aspects of the various reform
proposals are captured by the three options offered by the Advisory Council on So-
cial Security in 1996. Common to all three options is that they would permit either
public or private equity investment. This article discusses the economic risks in-
volved in public and private equity investments as a funding solution for Social
Security. To quantify the risks involved in equity investment, stochastic simulations
are based on the economic assumptions of the 1998 Trustees Report of Old Age and
Survivors Insurance and Disability Insurance in combination with different as-
sumptions about the rates of return on bonds and stocks. For public equity invest-
ment, financial market risk remains significant for at least 40 years. For individual
accounts, I find that the chance of doing worse than with Social Security or of
falling into poverty in retirement is generally high, yet varies with income level, gen-
der, family status, and employment history. In general, women, married workers
with dependent spouses, or workers with incomplete work histories fare worse than
men, single workers, or workers with complete work histories when compared ei-
ther to the current system or to the poverty line. © 2000 by the Association for Public
Policy and Management.
INTRODUCTION
Since the 1930s, Social Security has been the foremost retirement system for workers
and their families in the United States. However, recently the continuation of this
success story has been questioned. The Trustees of the Old Age and Survivors Insurance
and Disability Insurance (OASDI)—as Social Security is officially called—project in
their 1998 report that after 2032, revenues will cover only about 75 percent of benefits
allotted under current law if Social Security remains unchanged.
Ensuring Social Security’s financial soundness is a high priority because of its sheer
size as a government program, but also because of its importance for workers and
their families as income insurance, of a $362 billion total in 1997, the largest share
went to retirees, but 44 million workers and their families also depended on monthly
benefits. Particularly, low-income elderly Americans depend on Social Security’s
benefits, which provide on average 89 percent of the household income in the bottom
Journal of Policy Analysis and Management, Vol. 19, No. 2, 263–273 (2000)
©
2000 by the Association for Public Policy Analysis and Management
Published by John Wiley & Sons, Inc.

2
64 / Evaluating Market Risk of Equity Investment Proposals to Reform Social Security
quintile where the head of the household is 65 years or older. But that is not to say that
Social Security is unimportant for rich people: High income retirees over the age of 65
in the top quintile receive on average 29 percent of their income from Social Security.
To maintain a program on which so many elderly beneficiaries depend, various
solutions have been proposed, including some that would allow for public or private
equity investment [Advisory Council, 1997; GAO, 1997; Munnell and Balduzzi, 1998].
The favorable attention equity investment is currently receiving results from the
historically higher rate of return of stocks as compared to bonds. Particularly, the
prolonged bullish stock market since the 1980s has enticed policymakers to consider
equity investments more seriously. Essentially, those who favor equity investments
argue that if this higher rate of return continues, a significant part of the expected
shortfall could be financed.
Concerns over equity investment arise from the economic and political risks involved;
for instance, the government could become a major shareholder in U.S. companies.
A full discussion of the political risks involved is, however, beyond the scope of this
article, which will focus on financial market risks. One important aspect of economic
risks is that economic risks are socialized in the case of public equity investments,
while they are privatized in the case of individual accounts.
Risks that the trust funds or individual workers may incur in their retirement accounts
1
with equity investment cannot be ignored. Specifically, investments may underperform,
in which case the government would face unexpected funding needs in the future. Since
investors should theoretically be compensated for the greater risk of equity with higher
rates of return, the compounding of these returns should lower the overall risk over
time. Thus, over a long enough investment period, investors are more likely to approach
the average rate of return and to lower their risk. Even though there is a good chance
that in the long run stocks are virtually guaranteed to outperform bonds, the question is
how long the long run actually is. Leibowitz and Krasker [1988] show that over 30 years,
a stock portfolio underperforms a bond portfolio in 21 percent of the cases. Hence, both
the public and individuals may face significant risks despite a long investment horizon.
To gauge equity investments, the three proposals of the Report of the 1994-1996
Advisory Council on Social Security [1997] are used. Specifically, the expected rates
of return and their respective spreads—a standard measure of risk—are calculated
with the help of stochastic simulations. Then, the rate of return assumptions are
changed to show the sensitivity of the results. Finally, a number of probabilities are
computed, which could help policymakers decide whether a proposed solution imposes
too much risk on the national retirement system.
BACKGROUND
Stochastic simulations are used to imitate the market risks of equity investments. In
particular, the simulations are based on the income and outgo projections (the
intermediate cost projections) of the Trustees’ Report [OASDI, 1988]. Essentially, the
public can invest current assets and future surpluses in securities while individuals
can invest their expected income. The relevant question then becomes what rate of
return these assets will yield. Since nobody can accurately forecast the future, the
best we can do is to estimate a probability distribution of future outcomes. This is
done by 1,000 times randomly generating hypothetical returns for a 75-year investment
horizon. Then, each randomly produced 75-year set of hypothetical rates of return is
combined with the given assets and surplus figures of the trust funds, or—in the case
1
While Social Security provides retirement, disability, and survivorship benefits, I focus on retirement
benefits because equity investment proposals are restricted to this aspect of the program.

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 265
of individual accounts—with the expected earnings records of hypothetical workers.
With a large enough number of simulations, the distribution of outcomes should
approximate the future as long as the underlying assumptions hold. Thus, even such
outcomes as a constant rate of return, or long runs of above or below average rates of
return, are captured by the mean and the standard deviation of the sample.
Aside from the economic assumptions, investment schedules must be chosen that
cover a wide enough spectrum of policy proposals to be useful for the policy debate.
The options of the Advisory Council provide a range both diverse enough to cover
more recent proposals and realistic enough to include only proposals that are likely
to find political backing. The Advisory Council submitted three different options, the
so-called maintained benefits (MB), individual accounts (IA), and personal security
accounts (PSA) options. All three proposals allow for the option of equity investment
either by Social Security or by individuals. The main difference among these proposals
with respect to equity investment is that the MB option would socialize the risks
involved whereas the IA and PSA options would privatize the risks. In particular, the
MB proposal includes a suggestion to change the trust fund ’s investment strategy to
permit equity investment, such that a share of the current assets and future surpluses2
would be invested in equity until 40 percent of all assets were allocated in equity.
Furthermore, the IA and PSA options suggest diverting some portion of payroll taxes
into individual accounts while reducing Social Security’s benefits. Under the IA
proposal, three benefit cuts would be introduced: The increase in the normal retirement
age (NRA) from currently 65 to 67 would be accelerated, and it would increase with
longevity after 2011, the replacement ratios for middle- and high-wage workers would
be lowered, and the benefits computation period would increase from 35 to 38 years.
Additionally, workers would pay an extra 1.6 percent of payroll into their individual
accounts. Actuarial balance under the PSA proposal would be achieved by giving all
new workers the same monthly benefit, $410 in 1996 dollars; by establishing individual
accounts from contributions of 5 percent of payroll; and by levying an extra 1.52
percent of payroll as transition costs for the next 72 years.
The equity aspects of the three Advisory Council options are also included in a
number of more recent proposals. President Clinton’s own proposal calls for equity
investment by the trust fund. A proposal by the National Commission on Retirement
Policy includes an equity investment option similar to the IA proposal. And proposals
by Senators Daniel Moynihan and Bob Kerrey, or economists Martin Feldstein and
Andrew Samwick include features similar to those proposed in the PSA option, albeit
at much lower levels. President Clinton’s plan would allow for the trust fund’s equity
investment by giving part of the budget surplus to the trust funds, part of which
would be invested in private securities. The National Commission on Retirement
Policy proposes that 2.0 percentage points of the payroll tax should go to individual
accounts following the model of the Thrift Savings Plan for federal employees. This
proposal is similar to the IA option in allowing for individual accounts, but pools the
invested funds and restricts investment options to lower administrative costs. The
Moynihan–Kerrey plan would allow for a voluntary contribution of 1 percent of payroll
to individual accounts managed by financial institutions in return for a payroll tax
cut of 2 percentage point. Under the Feldstein–Samwick plan, workers would have
the option of contributing 2 percent of their earnings (up to the Social Security cap)
to personal retirement accounts, for which they would receive tax credits taken out
of budget surpluses. Upon retirement, a worker’s Social Security benefits would be
reduced by 75 cents for each dollar of benefits from the worker’s individual account.
2
The MB option also includes other revenue increases and benefits cuts. These would enhance trust fund
financing and hence affect the risk-return trade-off.

2
66 / Evaluating Market Risk of Equity Investment Proposals to Reform Social Security
To offset the loss of income to Social Security under the individual accounts option,
benefits would be cut as they are in the IA and PSA options.
While the following discussion about the risks associated with different forms of
equity investments apply in their specific form only to the three Advisory Council
options, the lessons that can be learned apply to all other proposals that include
some form of equity investments. In particular, the results of the stochastic simulations
apply only to the specific proposals, when it comes to actual dollar amounts. Also,
comparisons with Social Security or the poverty line apply directly only to the specific
proposals discussed in proposals similar to the Advisory Council’s proposals as most
such proposals are. The following results offer a general sense of the risks involved.
Furthermore, the risks involved in equity investments depend on the assumptions
made about rates of return and their standard deviations, as well as investment
horizons—aspects that apply to any investment proposal. In other words, the stochastic
simulations can be sufficiently generalized to a range of existing proposals.
OASDI TRUST FUND EQUITY INVESTMENT
The goal of the MB plan is to achieve a 40 percent equity allocation, either gradually or
rapidly. A rapid transition from bonds to stocks would allow the trust funds theoretically
to take advantage of the higher rate of return for the longest time possible. Following
three different investment strategies, a conservative, an intermediate and an aggressive
strategy, shows whether there is a difference in risk between gradual and rapid equity
allocation. Under the conservative strategy, initially 10 percent of the current trust fund
assets would be transferred into equity, and 25 percent of all future surpluses would be
used for equity allocations. The intermediate strategy would also include a 10 percent
initial allocation, but put 50 percent of all future surpluses into equity. The aggressive
strategy would begin with a 20 percent initial allocation, and would invest 50 percent
of all future surpluses into equities. Once the equity allocation ratio equalled to 40
percent, this ratio would be maintained until the trust fund was depleted.
By combining economic forecasts of the OASDI Trustee ’s Report with assumptions
about stock and bond returns, the investment risk for the trust funds can be gauged
by simulating 1,000 different possible outcomes. The program income and
expenditures figures are taken from the trustee’s report, while the bond and stock
returns can rest on different, and often contentious, assumptions. More specifically,
there are two sets of assumptions each for stock returns and bond returns: in both
cases, there are historical distributions, plus a set of lower stock returns more
consistent with other trustee assumptions or, in the case of bonds, higher bond returns
taken directly from the 1998 Trustees Report. Thus, depending on which of the four
sets of assumptions is chosen, the extra years before the trust funds are depleted,
which can be gained with equity investment, can vary greatly.
First, the most optimistic set of assumptions is chosen for equity and bond returns.
The bond returns are based on the trustees’ projections for long-range interest rates
with a mean of 6.3 percent and a standard deviation of 0.34 percent. The stock returns,
on the other hand, are based on historical returns since 1926 with a mean of 13.10
percent and a standard deviation of 20.18 percent.
On average, all equity investment strategies outperform the baseline scenario
Table 1), such that, depending on the strategy, additional trust fund assets can cover
(
the income shortfall for another 9 to 15 years. While in the baseline scenario, where
all funds are invested in government bonds, the funds are expected to run out by
2
031; with the conservative strategy it runs out by 2040; with the intermediate strategy
by 2044; and with the aggressive strategy by 2046. As expected, the greater mean rate

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 267
Table 1. The risk of trust fund equity investment.
Assumptions
Investment
Strategy*
Expected Trust Fund
Depletion Year
Market Risk†
Bond returns based
on 1998 Trustees Report;
stock returns based on
historical returns
I
II
III
IV
2032
2040
2044
2046
0.55
10.53
12.19
12.81
Bond returns based
on historical returns;
stock returns based on
historical returns
I
II
III
IV
2029
2037
2039
2041
1.21
8.84
10.40
11.23
Bond returns based
on 1998 Trustees Report;
lowered stock returns
I
II
III
IV
2032
2035
2037
2038
0.55
5.763
7.49
8.35
Bond returns based
on historical returns;
lowered stock returns
I
II
III
IV
2029
2033
2035
2036
1.21
4.78
6.15
6.82
*
Strategy I: all bonds strategy.
Strategy II (conservative): 10 percent initial equity investment; 25 percent of surplus in
equity.
Strategy III (intermediate): 10 percent initial investment; 50 percent of surplus in equity.
Strategy IV (aggressive): 20 percent initial equity investment; 50 percent of surplus in
equity.
†
Standard deviation in years.
of return from equity investment is traded off against an increase in risk, that is, a
higher standard deviation.
The additional 9 to 15 years of Social Security financing are based on a particular set
of assumptions that may not be realistic. Specifically, the assumptions of the trustees
report with respect to long-range interest rates are questionable since they contain
virtually no variability over the long run and since they are higher than the historical
averages. Instead, bond returns could be based on the historical distribution of interest
rates on government special issues—the current trust fund investment—from 1937 to
1
980 with a mean of 5.46 percent and a standard deviation of 3.17 percent.
With a lower mean bond rate, assets are depleted earlier with the baseline strategy
than under the trustees projections. Similarly, with all equity investment strategies
assets are expected to be depleted three to five years earlier than under the more
optimistic assumptions (Table 1). Despite the now-greater variance in bond returns,
investing only in bonds remains the least risky strategy.
Not only may the original assumptions about future bond returns be unrealistic,
but so may be the assumptions about future stock returns. A few researchers have
argued that, at least in the medium term, stock returns are likely to be below their
historical long-term average [Golob and Bishop, 1997]. In particular, Baker [1997]

2
68 / Evaluating Market Risk of Equity Investment Proposals to Reform Social Security
has pointed out that projections of future stock market rates of return based on past
performance are inconsistent with the other trustees’ economic assumptions. Instead,
a real rate of return closer to 4 percent may be more likely, which has historically
corresponded to a nominal mean rate of return of 10.92 percent and a standard
deviation of 17.07 percent.
The lower projections for stock returns in combination with the two sets of
assumptions for bond returns generate significantly earlier depletion dates than the
more optimistic stock return assumptions. The combination of lower stock returns
and the trustees’ assumptions for bond returns lets the trust funds become depleted 5
to 8 years earlier, while the combination with historical bond returns shortens the
average depletion date by 4 to 5 years (Table 1). So far, the focus has been on the
average performance of equity investments, which says nothing about their economic
risk. For policy purposes, the risk of equity investments is measured by three yardsticks.
First, policymakers may be interested in the chance that the trust funds will be depleted
before their expected depletion date. Second, the effect on government finances in
case of an unexpected and premature depletion should be discussed. Hence, the
probability that a shortfall of more than 1 percent of gross domestic product (GDP)
should be covered. Third, a baseline scenario may face equally a chance of being
depleted earlier than expected, so equity investments should be compared with the
current strategy of investing in bonds alone.
If equity investments are permitted, the chance of trust funds depletion in any
given year grows slowly around 2020. In contrast, without equity investment, the
same probability is about zero until about a decade later. Hence, with any equity
investment, policymakers may need to worry a decade earlier about the possibility
that the trust funds may be depleted prematurely.
The trust funds could reduce the depletion risk by choosing the most aggressive
strategy. Even then the chance that the trust funds are depleted 5 to 15 years earlier
than expected remains high (Table 1). Under the most optimistic assumptions, the
probability that in 2025—more than 15 years earlier than projected—no trust fund
assets will be left is about 3 percent. There is also a 2 percent chance under the most
pessimistic assumptions that the trust funds will be depleted in 2025—or about 10
years earlier than expected. Moving closer to the expected depletion date also raises
the chance of trust fund depletion, for example, in 2030 it is 11 to 13 percent, and in
2
035 it is 24 to 31 percent.
While these probabilities may seem small, their significance becomes clearer when
put in context. Let’s take bond and stock returns given by historical distributions.
Depending on the investment strategy, the expected depletion date lies between 2037
and 2041, while the chance that the trust funds will be depleted is about 4 percent in
2
025 and 14 to 17 percent in 2030. For 2025 and 2030, though, the 1998 trustees’
report projects a Social Security revenue shortfall of 2.42 percent and 3.05 percent of
GDP, respectively. If the trust funds were unexpectedly depleted, these shortfalls would
have to be covered. For 2030, the OMB [1999] is projecting essentially a balanced
3
primary budget (without Social Security). Thus, there is a 14 to 17 percent chance
that in 2030 the government will have to cover a unified budget deficit of 2.4 percent
or 3 percent of GDP by raising taxes, cutting expenditures, or taking out new debt,
when its finances are already constrained.
As another way of getting a sense of the fiscal risks involved, the probability of
covering a loss of more than 1 percent of GDP was calculated for the years 2025,
2
030, and 2035. The government could cover unexpected additional expenditures to
3
Discretionary spending grows with population growth [OMB, 1999, Table 2-2].

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 269
the tune of 1 percent of GDP or less, while anything bigger may pose large political
and economic obstacles. In 2025, the probability is 0.4 to 3.4 percent, while it increases
to 10 to 17 percent in 2030 and further to 22 to 41 percent in 2035. Interestingly, the
probabilities decline significantly as the aggressiveness of the equity investment
strategy increases. However, even with the aggressive strategy, the chances of covering
unexpectedly high losses several years before the expected depletion dates stay well
above 10 percent.
The next question is whether Social Security faces a lower risk if it continues to
invest in special government bonds. After more than 35 years, there is still a 13 to 15
percent chance that an all-bonds strategy outperforms any of the equity allocation
strategies under the first set of assumptions. Similarly, the probability of performing
worse than the baseline scenario remains between 6 percent and 15 percent for the
years 2025, 2030, and 2035. The risk of underperforming with equity investments
can be limited by following the most aggressive strategy. More aggressive strategies
have a smaller chance of underperforming the baseline scenario than more
4
conservative ones—at least over the long run. Though while aggressive strategies
lower the chance of performing worse than an all-bond strategy, this probability
remains positive for almost 40 years.
INDIVIDUAL ACCOUNTS UNDER THE IA OR PSA PROPOSALS
An important distinction is that individual accounts privatize the risks from equity
investments, while trust fund equity investment socializes the risks. That is, for the
equity investment aspects of Social Security, each worker would bear the entire risk
as individual accounts would not be affected by the redistributive mechanisms of
Social Security, while trust fund accounts would.
To evaluate retirement income security under the IA and PSA proposals, first the
expected terminal account balances are calculated, and then the benefits of each
5
option compared to the current benefits. Again, to give a sense of the market risks
involved, both the probability that the terminal account balances fall below 25 percent
of their expected value, and the chance that without additional retirement income, a
worker has an income below the poverty line are provided. With individual accounts,
market risks are obviously not the only investment risks. Individual risks, such as
unemployment, low life expectancy, or simply poor investment decisions, should be
considered. Most characteristics can be accounted for in a stylized manner and are
consequently included. However, poor investment decisions cannot be considered.
That is, in each case, the investment outcomes reported are the results for an average
investor. Individual results may vary more, especially under the PSA option, where
workers can choose from a range of options. To account for individual risk factors,
4
The most conservative strategy always performs better than more aggressive strategies in the short run,
while the ordering reverses between 2025 and 2030. This reversal in the ordering shows that the greater
rate of return eventually begins to outweigh the greater risk.
5
The focus here is solely on retirement benefits as under both options survivor and disability benefits are
linked to the basic retirement benefits levels. Because in both cases the currently mandated law benefits
are reduced, the comparison of survivorship and disability benefits would likely be less favorable for the
privatized plans than the comparison for retirement benefits.
6
Considering that it is unreasonable to assume that any worker will continuously have above-average
earnings, the figures for high earnings workers are “best-case scenarios.”
7
For the individual accounts options, I assume historical returns. By changing the assumptions, the
chances increase that partially privatized plans will do worse than Social Security or worse than poverty
thresholds. Also, privately managed accounts incur administrative costs not included here, which hence
result in an overstatement of the performance of PSAs.

2
70 / Evaluating Market Risk of Equity Investment Proposals to Reform Social Security
some hypothetical cases are used. In general, it is assumed that workers work at
6
three different earning levels and do not move up or down in earning. At the start of
the investment period, the worker decides to invest either all in bonds, all in stocks,
7
or half in bonds and the other half in stocks for the next 35 years. Moreover, gender,
family status, and work history are added to the discussion.
The risk involved in IA proposals if a worker works continuously for 35 years at
minimum, average, or high earnings—as defined by the 75-year forecasts of the 1998
Trustees’ Report—and contributes under either the IA or the PSA option grows with
equity investments. For instance, the simulation results show that the risk that after
3
5 years of contributions a worker has less than 25 percent is 0 percent with no
equity investment, 18 percent if half of all annual contributions, and 23 percent if all
contributions are invested in equity under the IA option (Table 2). The respective
chances increase to 21 percent and 26 percent under the PSA option. These
probabilities hold for all earning levels and income histories since they reflect market
risks over a 35-year investment horizon, independent of personal characteristics.
Clearly, workers should be willing to bear a greater risk if they are sure to fare
better than under the current system. To make the two proposals comparable with
Social Security, both benefits cuts and additional taxes, must be factored in, as workers
would contribute more to the new system and should therefore expect more in return.
After accounting for the proposed changes the benefits are scaled by (12.4)/(12.4+1.6)
for the IA option and by (12.4)/(12.4+1.52) for the PSA option. The chance of doing
worse with individual accounts than with Social Security is significant. Any equity
investment leaves a worker with a 19 to 34 percent chance of doing worse under the
IA plan, and with a 13 to 21 percent chance of underperforming under the PSA option.
Another way to look at market risk is to estimate the chance that a worker’s benefits
from either option falls below the poverty line (Table 3). Now, a worker ’s personal
characteristics seem to matter: Women fare worse than men, married workers fare
Table 2. The risk of equity investment in individual accounts.
Earnings Level
Investment
Strategy*
Probability that Terminal Account
Balance after 35 Years Is Less Than
25% of the Expected Value
IA Option (%)
PSA Option (%)
Minimum
Average
High
I
II
III
0.0
18.2
23.2
0.0
21.1
25.6
I
II
III
0.0
18.2
23.2
0.0
21.2
25.6
I
II
III
0.0
18.2
23.2
0.0
21.2
25.7
*
Strategy I: investment in bonds only.
Strategy II: 50 percent of annual contibutions each in stocks and bonds.
Strategy III: investment in stocks only.

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 271
Table 3. Probability of falling into poverty for worker with average earnings with IA
or PSA option (no other income assumed).
IA Option
Investment Strategy*
Personal Characteristics
I
II
III
Single male with complete work history
Single female with complete work history
Married male householder with complete work history
2.9
11.9
93.2
18.8
20.4
26.9
20.9
21.7
24.9
Single male with incomplete work history
Single female with incomplete work history
Married male householder with incomplete work history
1.2
6.1
94.7
20.4
21.8
28.5
22.6
23.3
26.7
PSA Option
Single male with complete work history
Single female with complete work history
Married male householder with complete work history
0.0
0.0
0.3
14.9
15.5
16.9
18.7
19.0
19.9
Single male with incomplete work history
Single female with incomplete work history
Married male householder with incomplete work history
0.0
0.0
0.0
16.8
17.3
18.2
20.6
20.9
21.4
*
Strategy I: investment in bonds only.
Strategy II: 50 percent of annual contibutions each in stocks and bonds.
Strategy III: investment in stocks only.
In the case of a married male householder, only retirement benefits are included.
worse than single workers, and workers with interrupted work histories fare worse
than workers who have continuously worked. Hence, male married workers with an
interrupted work history who have received average earnings for 25 years face a 29
percent chance of falling below the poverty line under the IA option and a 21 percent
chance under the PSA proposal—if they have no other sources of income.
Besides market risk, gender, family status, and work history are considered. Women
have a longer life expectancy than men and hence a longer retirement. Financing a
longer retirement, though, automatically poses a greater risk for a worker under the
IA option. Put differently, a worker whose only distinction is a longer life expectancy
would have to accumulate a greater terminal account balance to receive the same
annuity than an otherwise identical worker. Social Security automatically compensates
for different life expectancies, and each worker with the same earnings record receives
the same monthly benefit. Another feature of Social Security is its family benefits.
That is, Social Security provides additional benefits for dependents. Such an insurance
aspect is still incorporated in the basic benefit parts of the IA and PSA proposals, but
it does not exist for the individual accounts parts. Finally, a result of Social Security’s
redistributive features is that workers with incomplete work histories receive larger
benefits relative to their contributions than workers with complete work histories
who are otherwise identical.
Again, women, married workers, and workers with interrupted work histories have
a higher chance of doing worse than their counterparts with IAs when compared
with Social Security. For instance, for the IA option, the chance of doing worse is by

2
72 / Evaluating Market Risk of Equity Investment Proposals to Reform Social Security
0
.8 to 1.8 percentage points higher for single and by 1.8 to 4.5 percentage points for
married women, while under the PSA option, for married and single women the
chance of doing worse is 0.1 to 1.8 percentage points greater than for men—if an
equity investment option is chosen. Further, in comparison to single workers, the
chance of doing worse grows by 4 to 13 percent under the IA option and by 1 to 5
percent under the PSA option if the worker is married. To proxy an interruption in
one’s work life, the hypothetical worker stops working for 10 years after having worked
8
for 10 years. A single worker with such an interruption is likely to do worse in 20 to
2
3 percent of the cases for the IA option, and in 17 to 22 percent of the cases for the
PSA option. Workers with dependent spouses have a greater chance of doing worse
with the IA plan but not with the PSA option. With a dependent spouse, a worker’s
chance of doing worse grows by 4 to 15 percentage points.
9
CONCLUSION
A number of proposals to address the expected funding shortfall of Social Security
have included suggestions to divert some of workers’ contributions to private securities
markets. This article assesses the financial market risks involved in three of such
proposals—the MB, the IA, and the PSA options of the 1995–1996 Advisory Council
on Social Security.
How the OASDI trust funds would fare with equity investment depends on a few
factors. First, investing in equity may generate higher average returns, and thus a
longer financing horizon, but how many additional years of positive trust fund
assets depends largely on assumptions about stock market returns. Second, the
greater risk of equity investment is unlikely to disappear. Thus, a significant chance
remains that the trust funds may be depleted earlier than expected in the 1998
Trustees Report, which is the year 2032. Third, while the results suggest that if
equity investment is permitted for the trust funds, it may be best to follow an
aggressive strategy, such an aggressive strategy may have serious financial market
implications as inflows into equity markets could rise dramatically, thereby inflating
equity prices in the short run.
The results for IA options are more diverse because they depend on a worker’s
personal characteristics. Generally, the chance of falling into poverty with partially
privatized accounts is significant, namely above 10 percent with complete earnings
records and above 15 percent with interrupted earnings histories, if a worker has no
other retirement income. Further, women are more likely to fare worse with IAs than
men simply because of their greater life expectancy at age 65. Third, married workers
where only one spouse has an earnings record are slightly worse off under the IA plan
than single workers. Fourth, workers with incomplete work histories are more likely
to do better with Social Security than workers with complete earnings records.
I am grateful to Doug Fulton for invaluable technical assistance. I also would like to thank
David Smith, Dean Baker, Tom Palley, and two anonymous referees for comments on an earlier
draft of this paper. All remaining errors are entirely my responsibility.
CHRISTIAN E. WELLER is a Research Economist at the Economic Policy Institute,
Washington DC, and Senior Fellow at the Center for European Integration Studies,
University of Bonn.
8
Note that “the median number of years with zero earnings for workers turning 62 in 1993 was 4 for men
and 15 for women” [GAO, 1997].
9
This larger increase in the chance of doing poorly under the IA plan stems from the fact that the IA
option includes a gradual cut in spousal benefits, while the PSA proposal does not.

Evaluating Market Risk of Equity Investment Proposals to Reform Social Security / 273
REFERENCES
Advisory Council on Social Security. (1997). Report of the 1994–1996 Advisory Council on
Social Security. Washington, DC: U.S. Government Printing Office.
Baker, D. (1997). Saving Social Security with stocks: The promises don’t add up. A Twentieth
Century Fund/Economic Policy Institute Report. New York: Twentieth Century Fund.
GAO (General Accounting Office). (1997). Social Security reform—Implications for women’s
retirement income. Report to the Ranking Minority Member, Subcommittee on Social Secu-
rity, Committee on Ways and Means, House of Representative.
Golob, J.E., & Bishop, D.G. (1997). What long-run returns can investors expect from the stock
market? In Federal Reserve Bank of Kansas City economic review, third quarter 1997. Kan-
sas City, MO: Federal Reserve Bank.
Leibowitz, M.L., & Krasker, W.S. (1988). The persistence of risk: Stocks versus bonds over the
long run. Financial Analysts Journal, 44(6), 40.
Munnell, A.H., & Balduzzi, P. (1998). Investing the Social Security trust funds in equity. Wash-
ington, DC: Public Policy Institute of American Association of Retired Persons.
OASDI. (1998). Annual report of the Board of Trustees of the Federal Old-Age and Survivors
Insurance and Disability Insurance Trust Funds. Washington, DC: OASDI Board of Trustees.
OMB (Office of Management and Budget). (1999). Analytical perspectives: Budget of the United
States government, fiscal year 2000. Washington, DC: OMB.