Leonesio and Wixon 2004 — Increasing the Early Retirement Age Under Social Security Health, Work, and Financial Resources

Summary

A National Academy of Social Insurance (NASI) policy brief using 1990 Survey of Income and Program Participation (SIPP) data linked to Social Security Administration (SSA) administrative records (n=1,090, ages 62–64, 1990–92 reference period) to document the health and financial status of 62–64 year old early Old-Age and Survivors Insurance (OASI) retirees. The central finding: early retirement benefits function as an unofficial disability program for a significant minority of early retirees who are too impaired to work but ineligible for formal Disability Insurance (DI; no insured status) or Supplemental Security Income (SSI; assets above the $2,000 means test). Simulations show that eliminating early retirement benefits would shift only ~8% of health-impaired early retirees into DI or SSI, leaving the remainder with no safety net.

Key Claims

Health status of 62–64 year olds: Of 6.4 million people aged 62–64 in the early 1990s: 51% had no health problems; 22% had health problems but not severely disabled; 27% were severely disabled (substantial work impairment); 16% were simulated to meet the strict SSA disability test. Of the severely disabled, 39% received OASI and 40% received DI/SSI.
12% of early OASI retirees meet SSA disability criteria: About 12% of OASI beneficiaries aged 62–64 simulate meeting the medical eligibility standard for DI or SSI, yet receive neither — DI is blocked by insufficient work quarters; SSI is blocked by the $2,000 asset test.
Steep financial gradient within early retirees: Among OASI beneficiaries, comparing those with no health problems to those simulated to meet SSA disability criteria — poverty rate: 4% vs. 25%; median financial assets: $37,652 vs. $800; Social Security ≥50% of income: 32% vs. 58%; median lifetime indexed earnings: $9,318/yr vs. $2,332/yr.
Demographics of the gap population: The simulated SSA-disabled OASI subgroup is 79% female, disproportionately widowed/divorced/separated (53%), more likely to have less than 12 years of schooling (53%), and more likely to live alone (41%).
DI insured status blocks access: Only about 50% of women aged 62 in 1990 were insured for disability (vs. 80% of men). Women who spent years out of the paid labor force lack the 20-of-last-40-quarters work requirement for DI.
Eliminating OASI at 62–64 would not shift most to disability programs: Simulation shows only 5% of affected OASI beneficiaries would become eligible for SSI disability; 3% could access DI; 17% (533,000) with significant work limitations would have no program access at all.
Early retirement as de facto disability program: "About as many 62–64 year olds classified as severely disabled receive early retirement benefits as receive disability benefits from DI or SSI." Early retirement fills the gap left by DI's work-history requirement and SSI's asset test.

Concepts Introduced or Extended

Entities Mentioned

Quotes

"Social Security early retirement benefits serve as a substantial, albeit unofficial, disability program for some early retirees."

"About as many 62 to 64 year olds classified as severely disabled receive early retirement benefits as receive disability benefits from DI or SSI."

My Take

The core finding — that early retirement benefits backstop a population that formal disability programs fail to reach — is robust, though the magnitudes rely on SIPP-based disability simulation with known limitations (self-report measurement error, imperfect mapping from survey responses to SSA medical criteria). The 12% estimate is likely a conservative floor given that the simulation uses the same strict medical thresholds as the actual DI/SSI programs.

The most policy-relevant finding is the simulation result: raising the early retirement age without simultaneously reforming DI/SSI would leave roughly 17% of early retirees with significant work limitations and no fallback program. This is not a small rounding error — it represents hundreds of thousands of people. The DI insured-status gap and SSI asset test are not passive omissions; they are active structural features that deliberately exclude certain populations. Any early retirement age reform that does not address both would shift the incidence of policy failure onto the most vulnerable early retirees: low-earning women with interrupted work histories.

The paper also implies an understated point: to the extent that early retirement functions as a disability pathway, measured DI caseloads understate the true scale of health-related labor force exit among 62–64 year olds. This is a mirror image of Autor and Duggan's (2003) argument that DI expansion lowered measured unemployment — here, OASI expansion lowers measured DI receipt among the near-elderly disabled. See Early Retirement Safety Net Gap and Select and Ultimate Mortality Tables for the mortality and selection implications.