A number of proposals for Social Security reform have been introduced in recent years in response to concerns about Social Security's long-term solvency. Many of these reform plans call for individual accounts that would allow individuals to manage their own retirement savings. For example, the President's Commission to Strengthen Social Security (CSSS) proposed Social Security individual accounts (SSIAs) in December, 2001. The Commission's SSIAs were modeled in part on 401(k)-type plans like the Thrift Savings Plan (TSP) for federal employees. Other recent proposals also look to the TSP as a model for the administrative component of individual accounts.
The TSP is a successful voluntary defined contribution plan. It was designed as the individual savings tier of a 3-tiered approach to retirement income security for federal workers that includes Social Security as tier 1 and a defined benefit plan as tier 2. Although the TSP is the largest defined contribution plan in the country, with 2.9 million participants and approximately $100 billion in assets, it pales in size to the over 144 million workers who would be potential participants of Social Security Individual Accounts (SSIAs).
The feasibility of modeling individual accounts after the TSP has generally been accepted as an article of faith. Yet few analysts have probed the administrative implications of establishing SSIAs for employers, workers, the government, and the various financial service entities that would potentially be a part of this program. The author of this paper, Francis Cavanaugh, was one of the architects of the TSP and served as the first Executive Director and CEO of the Federal Retirement Thrift Investment Board that oversees the TSP.
Cavanaugh argues that a system of individual accounts cannot easily be adapted to the TSP format. In particular, SSIAs would not be cost effective because there would be too many low-income employees and too many very small businesses. More than 85 percent of businesses with fewer than 100 employees have no retirement plans. Small businesses generally do not have the financial resources or the personnel, payroll, or systems staffs to do what 401(k) employers now do. They could not be expected to assume the expenses or fiduciary responsibilities of 401(k) employers.
Per capita administrative expenses would average more than $300 a year, and even more for employees of very small companies, based on current market costs. For example, the CSSS's proposed annual contribution of 2 percent of pay (in its "basic" plan) to the average SSIA would be so small that more than 60 percent of the contribution would be required just to cover administrative expenses. Current average annual pay subject to Social Security taxes is about $26,000, so 2 percent would be about $520. The potential cost of more than $46 billion a year (155 million workers x more than $300 per account) would be more of an advantage to the 401(k) industry than the workers and their employers.
The prospects for future cost reductions are not good. The major 401(k) providers have been competing aggressively for many years to supply 401(k) plans to small businesses. They have achieved substantial economies of scale with state-of-the-art technology, but they still charge more than $3,000 a year for the average small business, which has fewer than ten employees. Thus major providers recommend against 401(k)s for businesses with fewer than ten employees, and some recommend against them for businesses with fewer than 20 or 25 employees. The creation of a new federal agency to administer SSIAs would add to total costs. Moreover, the error rates that are now tolerated by IRS and SSA would be totally unacceptable in a 401(k) environment of day-to-day investment transactions.
Reforming a part of Social Security to resemble 401(k)s is problematic because Social Security and 401(k)s are based on entirely different criteria. Social Security is mandatory - it is a tax; 401(k)s are voluntary, for both the employer and the employee. Social Security is an insurance program; 401(k)s are investment programs. Social Security depends on the government to absorb inflation and market risks; 401(k)s shift those risks to individual investors. Social Security is a defined benefit program; 401(k)s are defined contribution programs. Social Security is a safety net for lower-income people; 401(k)s are a major tax subsidy for higher-income people. Social Security is sponsored and administered by the government; 401(k)s are sponsored and administered by employers. As a result, issues that are not a problem for the 401(k) system — such as the size of contributions, administrative expenses, and feasibility for small businesses — would cause serious problems for Social Security.
These findings contribute important information to the discussion of Social Security individual accounts. In particular, Mr. Cavanaugh casts doubt on the viability of using the TSP as a model for individual accounts. At the very least, this paper underscores the importance of examining with great care the administrative structure and costs of the various proposals for individual accounts.
Written by Francis X. Cavanaugh
Laurel Beedon and Alison Shelton, Project Managers, AARP Public Policy Institute
May be copied only for noncommercial purposes and with attribution; permission required for all other purposes.
Public Policy Institute, AARP, 601 E Street, NW, Washington, DC 20049