Adding individual accounts to Social Security could be the largest undertaking in the history of the U.S. financial market, and no system to date has the capacity to administer such a system. The number of workers currently covered by Social Security—the largest single entitlement program in the nation—is at least four times higher than the combined number of all tax-favored employment-based retirement accounts in the United States, which are administered by hundreds of entities.
Direct comparisons between employment-based retirement savings plans and Social Security reform are tenuous at best. Social Security covers workers and businesses that are disproportionately excluded from employment-based plans. Because of these differences, a system of individual Social Security accounts would be more difficult to administer than employment-based plans, and total administrative expenses would be larger relative to benefits.
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. Inherent in the “privatization” debate is generally the presumption that IA benefits would be based on cash contributions and investment returns. The current credit-based system tolerates small errors in wage reporting, because they rarely affect benefits. But every dollar counts in a cash-based IA system. To ensure that benefits are properly provided, an IA system would require more regulation, oversight, and error reconciliation than the current Social Security program.
Social Security individual accounts cannot be administered like 401(k) plans without adding significant employer burdens—especially on small businesses. Under the current wage reporting and tax collection process, it would take at least 7–19 months for every dollar contributed to an individual’s account to be sorted out from aggregate payments and credited to his or her IA. This 7–19 month “float period” could result in substantial benefit losses over time. Options for preventing such losses involve difficult trade-offs, such as increased government responsibility, increased complexity, greater employer burdens, and/or investment restrictions for beneficiaries.
If legally considered personal property, the IAs of married participants could pose significant administrative challenges. Social Security today must obtain proof of marriage only at the time spousal benefits are claimed. But some IA proposals would require contributions to be split between spouses’ individual accounts, requiring records on participants’ marital status to be continuously updated to ensure that contributions are correctly directed. Also, dealing with claims on individual account contributions in divorce cases could place IA record keepers in the middle of spousal property disputes.
The current body of knowledge is too uncertain, and 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. Uncertainty exists over how IA proposals would address key policy areas affecting administrative cost and complexity, how administrative costs operate in the current employer-sponsored retirement arena, and how lessons from the employment-based system apply to Social Security reform.
Individual account benefits would be highly sensitive to administrative costs, according to results using SSASIM. Workers born in 1976 and 2026 would receive 40 percent to 42 percent lower IA benefits under high administrative cost assumptions than under low-cost assumptions, indicating that additional research on administrative costs is essential to assessing how—or whether—IAs could produce meaningful retirement benefits.
