Individual Social Security Accounts: Administrative Issues

For all the political and policy discussion over creating individual accounts (IAs) within Social Security, the debate so far has virtually ignored any specific considerations about the practical issues of how to administer such accounts. Any discussion of whether to create individual accounts must also address the basic but critical questions of how they would work: Who would run them? What would they cost? Logistically, in what form are they possible? This EBRI Issue Brief/Special Report provides an update to the November 1998 EBRI Issue Brief, which presented an overview of the most salient administrative issues facing the current Social Security reform debate-issues that challenge policymakers to think carefully about how their proposals could be implemented, in order to achieve their policy goals.

As this EBRI Issue Brief/Special Report discusses, there is a way to design a system of individual accounts that could potentially be administered in a cost-effective and timely way-but for a variety of inescapable reasons, that system most likely would bear little or no resemblance to a modern 401(k) plan. If a typical Internet-based 401(k) with easy access to account information and investment options can be described as the "Porsche" of retirement savings plans, then the public should realize that a workable, cost-effective individual account within Social Security will most likely look like a "Model T":

• The average 401(k) plan offers nine actively managed investment choices ("Porsche" plans offer virtually unlimited options through mutual fund "windows" and self-directed brokerage accounts); a startup universal IA system would offer a very limited number of index investment options-probably just one initially (the Federal Thrift Savings Plan offers only five).
• 401(k) plans typically offer daily access through the telephone, and "Porsche" plans offer round-the-clock Internet-based self-management with immediate access to account information, updated daily; some-thing closer to an annual account statement would be likely for a startup universal IA system.
• 401(k) plans typically offer participants loans or hardship withdrawals from their accounts, with "Porsche" plans providing "do-it-yourself" loans over the Internet; a start-up universal IA system likely would find it impossible to offer either.
• Workers' 401(k) contributions typically come out of every paycheck, with rapid crediting to investment accounts; a start-up IA system tied to Social Security would involve "bulk" contributions, with annual reporting of contributions to the Social Security Administration at the worker level, and crediting as much as 18 months later as the paper is processed. "Porsche" 401(k) plans do both contributions and allocations on a fully automated basis every pay period; a start-up IA system could not, since the majority of U.S. employers file with the government on paper.
• 401(k) plans allow participants to modify their contributions regularly, with "Porsche" plans allowing it round-the-clock over the Internet for next-pay-period implementation; a start-up IA system more likely would allow this once a year (per employer) when employees fill out their withholding form.
• "Porsche" 401(k) plans rely on employers and administrators being completely automated, with computer interface of all data; a start-up IA system would have to allow employers to continue using pen-and-paper reports-as most currently do-if high administrative costs are to be avoided (especially for small businesses).
• "Porsche" 401(k) plans offer proactive, personalized participant education and investment advice; a start-up IA system would likely offer only generic education materials.

The issues and options in administering IAs raise concerns that cut across ideology. The object of this report is neither to dissuade the advocates nor support the critics of individual accounts. Rather, it is to bring practical considerations to a political debate that must ultimately deal with the pragmatic challenges of designing IAs that would not be too complex for participants to understand, too burdensome for small employers to administratively support, too difficult for a record keeper to administer, or too expensive for low- and moderate-income participants to afford. The major findings in this analysis include:

• Adding individual accounts to Social Security could be the largest undertaking in the history of the U.S. financial market, and no system currently exists that has the capacity to administer such a system
• Direct comparisons between employment-based retirement savings plans and Social Security reform options are difficult at best.
• Credit-based systems such as the current Social Security program are less difficult to administer than cash-based systems, which must account for every dollar.
• Social Security individual accounts cannot be administered like 401(k) plans without adding significant employer burdens-especially on small businesses.
• If legally considered personal property, the IAs of married participants could pose significant administrative challenges.
• The current body of knowledge is too uncertain, and in general the proposals to date are too vague, to make an objective estimate of how much an IA system would cost to administer or whether it would succeed in accomplishing its policy goals.
• Individual account benefits would be highly sensitive to administrative costs, according to results using SSASIM.