Testimony:  Comparing Social Security Reform Proposals 
Jeff Lemieux

For the Senate Special Committee on Aging
June 15, 2004

Summary:  Social Security reforms have the potential to help future generations avoid large tax increases or benefit cuts, and to enable lower-income workers to more easily accumulate savings and wealth.  However, the hyper-politization of Social Security has dimmed the prospects for a constructive discussion.  To cut through the political noise and cultivate some bipartisanship, we need accessible, unbiased explanations of how reform proposals would impact the budget and affect workers.  Like government in general, I believe Social Security should be small but progressive.  Legislators should consider Social Security reform proposals that would restrain federal spending while also helping lower-income workers.

Outline:
A Need for Clear, Evenhanded Analysis
Four Criteria for Evaluating Reform Proposals
  Budget Impact
  Degree of Progressivity
  Wealth-Building Potential
  Use of Gimmicks or General Revenue Transfers
Comparison of Social Security Trustees' and New CBO Projections


Thank you, Senator Craig and members of the committee, for inviting me.  My name is Jeff Lemieux, and I’m the Executive Director of Centrists.Org.  Our mission is to help policymakers develop ideas that could achieve lasting bipartisan support on some of the toughest national issues -- health care, budget deficits, social security, and so on.

My comments will focus on how to evaluate Social Security reform proposals.  I will refer to several charts and tables, which are also printed in a supplemental handout.

A Need for Clear, Evenhanded Analysis

There are a few glimmers of bipartisanship and responsibility on Social Security reform. 

Representatives Jim Kolbe and Charlie Stenholm have a new bipartisan bill, and Democratic Rep. Harold Ford recently expressed conditional interest in Republican Senator Lindsey Graham's reform proposal.  Both proposals include moderately sized, progressive personal accounts and attempt to "pre-fund" a portion of Social Security benefits. 

Economists Peter Orszag and Peter Diamond have offered a responsible reform proposal that is not based on personal accounts.  As a result, their proposal does not require transition financing.

Unfortunately, these constructive proposals are the exception in the Social Security debate.  Most of the political discussion takes place at the much lower level of partisan name-calling and opportunism, scare tactics, or unrealistic promises.  

Because the debate is so shrill and simplistic, the most popular Social Security proposal is the "do nothing" approach.

At the other extreme is the "free-lunch" privatization idea, which would create very large personal accounts out of current payroll taxes.

I am concerned that legislators might not have a clear picture of either the budgetary or distributional consequences of these more extreme proposals. 

Evasive or deceptive proposals imply that tough choices are not needed.  Until they are more thoroughly analyzed, these sorts of proposals can seem too good to be true.

Part of the problem is that the issue of Social Security reform is so politically sensitive that "do-nothing" advocates and "free-lunch" proponents sometimes feel justified in exaggerating their claims, or in attempting to create a frame of analysis that obscures the central facts.   

A related problem is the hyper-politization of policy analysis in Washington.  Even the most bland or technical analysis can become fodder for partisan political propaganda if its implications aren't spelled out very clearly.

The data and comparisons below are mostly cobbled together from technical reports and tables from the Social Security Administration's Office of the Actuary (OACT).  OACT's analyses are highly professional, timely, and helpful.  However, because they are so technical, they can be prone to misinterpretation or false claims.

Four Criteria for Evaluating Reform Proposals

Social Security reforms should be evaluated on four criteria, in a descending order of importance: 

(1) impact on the budget,
(2) degree of progressivity,
(3) opportunities for wealth creation, and
(4) the presence or absence of gimmicks. 

1. Impact on the Budget.  The first reason for reforming Social Security is that its costs would otherwise raise government spending by about 2 percentage points of gross domestic product (GDP). Although Social Security costs are likely to be exceeded by both Medicare spending and interest expenses as the baby boomers retire (presuming the deficits and national debt are permitted to grow on their current trajectory), those extra 2 percentage points of GDP represent an important part of the larger entitlement problem (see Figure 1). 

Figure 1.



Another way to think about Social Security's future budget problem is to compare the program's benefit obligations with its dedicated tax revenues (mostly payroll taxes).  This measure is not perfect, because it's impossible to really force the federal government to use particular revenues for unique purposes.  However, it is a very common and logical approach.

Using long-term projections from the 2004 Social Security Trustees Report, Figure 2 shows that Social Security costs are expected to rise from 4.3 percent of GDP in 2010 to 6.3 percent by 2030.  Meanwhile, tax revenues dedicated for Social Security are projected to remain roughly flat at about 5 percent of GDP over the next 30 years.

Figure 2.



The difference between these costs and dedicated revenues is Social Security's effective cash-flow surplus or deficit.  Because benefit costs are currently lower than revenues, Social Security is running a surplus.  However, Social Security's current surplus is expected to disappear and become a deficit after 2010 as the baby boom generation starts to retire in large numbers.  After 2030, the Social Security deficit is projected to be between 1.5 and 2 percent of GDP (see Figure 3).

Figure 3.



By this measure, to fully solve Social Security's budget problem, a reform proposal must close the 1.5 to 2 percent of GDP gap between revenues and costs in the long run, without (1) creating overly high or unmanageable transition costs over the next 20 years, and (2) relying on financing gimmicks or "one-sided bets," such as overly optimistic assumptions about investment returns on personal accounts, unlikely claims about economic gains or increases in national savings, or absolute guarantees that Social Security benefits would never be reduced.

Consider five alternative approaches:

o  The "Do Nothing" Approach
o  The Diamond-Orszag Proposal
o  The Lindsey Graham Proposal
o  The Kolbe-Stenholm Bill in the House
o  A "Free-Lunch" Proposal (The Ferrara Plan)

Table 1 shows the projected costs and revenues of Social Security under each of these five alternatives.

---

Table 1.

Social Security Costs and Reform Options
(as a percent of GDP)
Selected Years		2010	2020	2030	2040	2050	2060
Baseline Benefit Costs		4.3	5.3	6.3	6.5	6.5	6.5
"Dedicated" Revenues		5.0	5.0	5.0	4.9	4.8	4.7
Gap		-0.7	0.3	1.3	1.6	1.7	1.8
Reform Options:
Sen. Lindsey Graham		1.1	0.9	0.3	-0.5	-1.3	-2.0
Kolbe-Stenholm		0.6	0.2	-0.2	-0.7	-1.2	-1.5
Peter Ferrara		2.4	2.2	1.3	0.2	-1.2	-2.0
Diamond-Orszag		-0.3	-0.3	-0.5	-0.8	-1.1	-1.5
Misc: Tax Max at $140k		-0.3	-0.3	-0.3	-0.2	-0.2	-0.2

Source:  Centrists.Org
Note:  Date from the 2004 Trustees Report and other reports from Social Security's Office of the Actuary.

---

By maintaining the status quo, the do-nothing approach implies a permanent Social Security deficit of between 1.5 and 2 percent of GDP, and federal deficits that are roughly 2 percentage points of GDP higher than today's.

The other four proposals would successfully close Social Security's funding gap in the long run.  However, the Graham proposal includes significant transition costs, which would otherwise burden current taxpayers. 

The Kolbe-Stenholm proposal does a better job of limiting and "paying for" its transition costs, although it's long-run savings are a little smaller.

The Ferrara proposal contains huge, economically untenable transition costs -- much higher than most other reform proposals. 

The Diamond-Orszag proposal, which mostly uses tax increases to solve Social Security's funding problem, does not have transition costs.  The budgetary downside to that proposal, however, is that its tax increases, though gradual, are permanent.

Figure 4 shows the cost of doing nothing vs. the cost of the Graham and Ferrara plans.

Figure 4.



Figure 5 shows the long-term impact of the Kolbe-Stenholm proposal on the Social Security operating deficit.

Figure 5.



Figure 6 shows the revenue increases included as part of the Diamond-Orszag proposal.

Figure 6. 



2. Degree of Progressivity.  If it were a private pension system, Social Security benefits would be calculated based on a one-to-one relationship to payroll taxes paid.  That is, a worker who paid twice as much in payroll taxes as another worker would receive twice the benefit.

However, Social Security is not a private pension system; it is a public social insurance system.  As social insurance, its benefits are generally proportionate to payroll taxes, but not in a one-to-one relationship. 

The purpose of these deviations from a strict one-to-one relationship between taxes and benefits is social:  to prevent poverty among the elderly and disabled workers, and certain of their family members. 

I believe a key criterion for Social Security reform -- and a prerequisite for bipartisanship -- is to maintain or enhance the program's progressivity. 

However, there isn't much official data on how the various reform proposals would stack up against the current degree of progressivity in the Social Security system.

I propose that in addition to requiring side-by-side tables on the cost and savings from Social Security proposals, Congress should insist on simple tables that illustrate the distributional consequences -- the degree of progressivity or regressivity -- as well.

For example, Table 2 shows the simulated impact of Senator Graham's proposal on workers who earned various amounts during their pre-retirement years.  The Graham plan would be somewhat more progressive than current law, with lower-income workers getting relatively higher benefits, and high-income workers receiving relatively lower benefits.

---

Table 2.

Simulation of Sen. Lindsey Graham Social Security
Proposal on Married Workers Retiring at NRA in 2034
Assuming Normal (60% Stock, 40% Bond) Investment Returns
(all amounts in 2003 dollars)
Average	Average	Average	Annual
Annual	Annual	Annual	Earnings
Earnings	Earnings	Earnings	Above
(each spouse)	(each spouse)	(each spouse)	(each spouse)
$16,600	$34,700	$55,600	$87,000
Current Law Monthly Benefit Promised (Not Funded)
860	1,417	1,879	2,294
Proposed Law Defined Benefit
590	822	1,180	1,505
Benefit from Individual Account (Basic Contribution)
232	381	417	415
Range of Benefit from Matched Contributions \a
80	6	0	0
414	29	0	0
Range of Proposed Law Benefit (Total)
902	1,209	1,597	1,920
1,236	1,232	1,597	1,920
Source: Centrists.Org.
Notes: Assumes default investment returns based on 60%
stock, 40% "safe" long-term bond return. Based on
couples with equal work careers and earnings.
Couples have typical earnings profile (except
highest-earning couple assumed to have earnings
at or above the taxable maximum in all years). It is
important to remember that comparisons with currently
promised benefits may be misleading if the government
would not be able to fund those promises in the future.
Calculations do not include income from unmatched
voluntary contributions. NRA = normal retirement age.
\a Low range of benefit from matched contribution assumes
low-income worker voluntarily saves $101 per year and
receives gov't match of $150. High range assumes low-
income worker voluntarily saves $801 per year and
receives maximum gov't match or $500.

---

Table 3 shows a similar distributional comparison of the Graham plan, using data from an alternative source, and for a different year and family structure.  Table 3 also includes a hypothetical minimum wage worker, which is especially helpful in illustrating the improvement in progressivity under this plan.

The estimates in Tables 2 and 3 are admittedly sketchy.  They were computed based on unpublished data, and cannot be easily replicated.  I apologize for that.

However, the results seem reasonable, and further analyses by OACT, the Congressional Budget Office (CBO), or the Congressional Research Service will probably show similar patterns.  Hopefully, the official estimating agencies will publish tables like Tables 2 and 3 as a matter of course, so that lawmakers can quickly view the distributional impacts of the various plans.

---

Table 3. 

Simulation of Sen. Lindsey Graham Social Security
Proposal on Single Workers Retiring at Age 62 in 2041
Assuming Normal (60% Stock, 40% Bond) Investment Returns
(all amounts in 2003 dollars)
Minimum	Low	Average	High	Annual
Wage	Wage	Wage	Wage	Earnings
Earner	Earner	Earner	Earner	Above
$10,712	$16,600	$34,700	$55,600	$87,000
Current Law Monthly Benefit Promised (Not Funded)
636	774	1,285	1,699	2,052
Proposed Defined Benefit and Basic Individual Account Benefit
684	723	1,115	1,407	1,655
Range of Benefit from Matched Contributions \a
151	151	12	0	0
783	784	64	0	0
Range of Proposed Law Benefit (Total)
835	874	1,127	1,407	1,655
1,467	1,507	1,179	1,407	1,655
Source: Centrists.Org.
Notes: Assumes default investment returns based on 60%
stock, 40% "safe" long-term bond return. Based on
single workers with typical earnings profile (except.
highest-earning worker assumed to have earnings
at or above the taxable maximum in all years, and min.
wage worker assumed same wage all years). It is
important to remember that comparisons with currently
promised benefits may be misleading if the government
would not be able to fund those promises in the future.
Calculations do not include income from unmatched
voluntary contributions.
\a Low range of benefit from matched contribution assumes
low-income worker voluntarily saves $101 per year and
receives gov't match of $150. High range assumes low-
income worker voluntarily saves $801 per year and
receives maximum gov't match or $500.

---

Although I have no data to show the committee -- not even sketchy estimates like Tables 2 and 3 -- I believe both the Kolbe-Stenholm and Diamond Orszag plans are considerably more progressive than current law.

However, my hunch is that the Ferrara proposal would be slightly less progressive than current law.

3. Opportunities for Wealth Creation.  I believe that Social Security reform could be used as a  lever to help many lower-income workers find easier ways to save and accumulate a modicum of financial wealth.

Too many Americans approach retirement without sufficient financial assets.  Many own a home, but possess few financial assets.  Some reach retirement age without assets of any kind, and are therefore completely dependent on Social Security and (possibly) some welfare programs.

Over the last two decades, the availability of generous defined benefit pensions (a fixed pension based on years of service and wages at a given firm) has declined.  It is no longer common for workers to stay with one employer for 20 or 30 years and retire with a pension that replaces a decent percentage of their income.

A large percentage of workers now have access to 401k savings programs at work, and many people have established Individual Retirement Accounts (IRAs) or an alphabet soup of other tax-favored savings programs.  However, some workers have used these retirement savings programs instead to fund education or housing or consumption expenses, sometimes paying a steep tax penalty as a result.

Sufficient retirement income is a moral and social imperative -- we simply don't want seniors to be desperately poor. 

However, asset building is also extremely important.  Income support is not enough.  In the U.S. capitalist system, wealth allows people to take risks, to launch businesses or educate children.  Without some wealth, families can't make the investments necessary to get ahead.  With some assets, people can borrow; without assets, it is hard to get that first loan.

Personal accounts that can be held as assets in a family are an important part of Social Security reform.  Accounts that are notionally owned by workers, but which are not actually under workers' control and which must be distributed in full when a worker retires do not pass the test of wealth creation.

4. Gimmicks, "Leverage," and Exotic Benefit Guarantees.  Proponents of Social Security personal accounts should use reasonable assumptions about the rates of return workers are likely to achieve, and should not promote the accounts as likely to lead to a large increase in national savings and investment.

Riskier investments earn higher returns in the long run precisely because they are risky -- the expected value of a risky investment is higher, but the possible deviations from the expected returns are also greater.  Risky private investments offer great opportunities for gains, but can also lose money.

Some proponents of Social Security private accounts insist that because private investments will have a high rate of return, the accounts are essentially "free money," creating national wealth that would not otherwise have been created.

In particular, proponents of large Social Security accounts claim that rates of return on those accounts will be so high that the government can be assured that all workers will be better off, and can guarantee the current Social Security promised level of benefits without risk.

These arguments are largely misplaced.  Even if workers decided to make risky investments in their Social Security accounts, there is no reason to believe that Americans' overall tolerance for risk will increase.  Investors may simply readjust their outside portfolios toward less risky investments, and overall national rates of return and levels of investment would be about the same.

An economist's maxim is "beware the free lunch," and Social Security calculations that sound too good to be true are usually based on the tenuous assumption that Social Security reform will dramatically increase national savings and investment, and raise the national tolerance for risk.  Proponents of Social Security accounts should not depend on such "leverage" in their calculations of the likely impact of reforms.

Social Security reform could spark increases in national savings, if the funding for personal accounts is at least partially "paid for" with tax increases or spending cuts, and if workers believe the funds in their Social Security accounts represent a replacement for benefit cuts and therefore do not adjust their outside levels of savings.

But the vague promise of higher national savings and returns on investment is not sufficient to create elaborate guarantees that all workers will be better off under reform.  The government should not be on the hook to compensate workers for investments gone bad, either within or outside the Social Security system.  Taxpayers should not be asked to take that risk, and if they did, we could be sure that investors would be less careful -- and possibly even reckless -- with their portfolio allocations.

A final thought on inflated claims and guarantees:  If these sorts of benefit guarantees were funded by the private sector -- which presumably has both the ability and the incentive to price them correctly -- the costs would probably be significantly higher than those indicated in standard actuarial analyses.

Comparison of Social Security Trustees' and New CBO Projections

Yesterday, the Congressional Budget Office (CBO) released a new report on Social Security's long-term outlook.  CBO's new cost projections are slightly lower than the Social Security Trustees Report within the 2020 -  2060 period, and CBO's outlook for revenues is slightly higher after 2040 (see Figure 7.)

Figure 7.



However, Figure 7 shows that CBO's new model is really telling us a very similar story:  Social Security benefits will jump by about 2 percentage points of GDP over the next 25 years, while revenues stay roughly flat.  For policymaking purposes, there is little real difference between CBO's new overall cost and revenue outlook, and that of the Social Security actuaries. 

Figure 8 shows how CBO's projection of Social Security's long-term surplus or deficit stacks up against the Social Security actuaries' estimate.  The gap between projected benefit costs and revenues is slightly lower in CBO's estimate, but the differences are not very large, and the overall conclusion about Social Security's financial situation shouldn't be changed. 

Table 4 (below) details how CBO's projections compare with those of the Trustees in ten-year increments between 2010 and 2060. 

Figure 8.

---

Table 4.

Social Security Costs and Revenues -- CBO vs. Trustees
(as a percent of GDP)
Selected Years		2010	2020	2030	2040	2050	2060
CBO (June 2004)
Benefit Costs		4.4	5.2	6.1	6.2	6.3	6.4
"Dedicated" Revenues		5.0	5.1	5.1	5.0	5.0	5.0
Gap		-0.6	0.1	1.0	1.2	1.3	1.5
2004 Trustees Report
Benefit Costs		4.3	5.3	6.3	6.5	6.5	6.5
"Dedicated" Revenues		5.0	5.0	5.0	4.9	4.8	4.7
Gap		-0.7	0.3	1.3	1.6	1.7	1.8
Source: Centrists.Org
Notes: Data from 2004 Social Security Trustees Report and the Congressional Budget Office (CBO)
"The Outlook for Social Security" (June 2004). Figures may not sum to totals because of rounding.

---

References:

Centrists.Org Transcript:  Addressing Greenspan's Challenge on Social Security and Savings:  Views from an Emerging Generation of Political Leaders
revised 04.02.2004 ||  Transcript of a discussion on Social Security reform held on March 25, 2004 with Senator Lindsey Graham and Rep. Harold Ford, moderated by Morton Kondracke of Roll Call.

Centrists.Org The Kolbe-Stenholm Social Security Reform Plan
revised 02.14.2004  ||  The new Social Security reform plan proposed by Reps. Jim Kolbe and Charlie Stenholm combines fiscal responsibility with the wealth-building potential of personal Social Security retirement accounts.

Centrists.Org Budgetary Effects of the Diamond-Orszag Social Security Proposal
revised 12.31.2003 ||  This proposal would resolve Social Security's funding shortfall directly, mostly by raising taxes.  The plan is fully "paid-for," without gimmicks or wishful assumptions.  The downside is that the tax increases would be permanent, not temporary. 

Centrists.Org Unfunded Transition Costs of the Ferrara Social Security Proposal
revised 12.10.2003 ||  Anti-tax activist Peter Ferrara's Social Security reform plan includes large personal accounts and a guarantee that benefits would not be reduced.  What's the catch?  The enormous unfunded transition costs.

Centrists.Org A Preliminary Analysis of Sen. Graham's Social Security Proposal
revised 12.14.2003 ||  The reform proposal introduced by Sen. Lindsey Graham of South Carolina this morning is serious and worthy of bipartisan consideration.  

Centrists.Org Raising the Cap on Payroll Taxes Doesn't Solve the Social Security Problem
11.17.2003 ||  Raising the cap on annual earnings subject to the payroll tax ($87,000 in 2003) would simply defer Social Security's cash-flow problem by a few years -- it isn't a permanent solution.