Summary
Working paper version of the foundational Autor-Duggan argument: post-1984 disability insurance (DI) growth is explained by two forces — the 1984 liberalization of the determination process and a mechanical rise in the earnings replacement rate for low-skill workers driven by wage inequality interacting with the progressive primary insurance amount (PIA) formula. Together these caused DI/Supplemental Security Income (SSI) rolls to more than double from 3.8 to 7.7 million (1984–2000). A key contribution is the "conditional applicants" concept and the unemployment-suppression counterfactual: measured U.S. unemployment would have been ~0.64–0.65 pp higher in 1998 absent the 1984 liberalization.
Key Claims
- Between 1984 and 2000, non-elderly adults receiving DI/SSI rose from 3.1% to 5.3%; absolute recipients more than doubled from 3.8 to 7.7 million.
- Replacement rate mechanism: The progressive PIA formula (average indexed monthly earnings, AIME → PIA) is indexed to the national average wage; because low-skill wages grew below average, the share of earnings replaced rose sharply for low-wage workers without any legislative change. Male 10th-percentile replacement rate: 53% (1979) → 63% (1998); 90th-percentile workers: no change.
- Conditional applicants: Workers who prefer employment when employed but apply for DI upon job loss. This is explicitly named and theorized here. Model shows that rising program generosity and worsening labor market conditions expand this group super-additively.
- Program liberalization: The 1984 Reform Act reversed a late-1970s clampdown; awards to under-40s more than doubled (1979–1993); mental/musculoskeletal share grew 60% (1983–1998); neoplasm/circulatory shares fell 40%; male:female beneficiary ratio fell from 2.06 to 1.35 (1984–1999).
- Post-reform shock responsiveness: After 1984, DI application and recipiency rates became 2–3× more responsive to plausibly exogenous labor demand shocks; high school dropout males increasingly exited the labor force rather than entering unemployment after job loss.
- LFP estimates: DI retrenchment (1979–84) raised male high school dropout labor force participation (LFP) by ~4.7–5 pp in high-retrenchment states; the post-1984 liberalization lowered it by a comparable amount. No comparable relationship for more educated groups.
- Unemployment counterfactual: 1 pp rise in state DI recipiency → −3.2 pp male high school dropout LFP (ordinary least squares [OLS]/instrumental variables [IV]). Combining with the industry demand shock faced by the typical 1984 high school dropout: DI liberalization reduced measured aggregate unemployment by ~0.64–0.65 pp in 1998. By 1999, ~60% of non-elderly male high school dropout non-participants were receiving disability benefits.
Concepts Introduced or Extended
Entities Mentioned
Quotes
"We refer to these individuals as 'conditional applicants.'" (p. 2)
"Our calculations suggest that the aggregate U.S. unemployment rate would at present be approximately 0.65 percentage points higher were it not for the liberalization of the disability program in 1984." (p. 4)
My Take
This Center for Retirement Research (CRR) working paper is the direct precursor to the 2003 Quarterly Journal of Economics (QJE) paper and contains the same core argument. Its main independent value is that it clearly coins the term "conditional applicants" and explicitly frames the unemployment-suppression counterfactual. The replacement rate tables here use slightly different base years than the 2003 published version (1979/1998 here vs. 1979/1999 there). Cite the 2003 QJE for the published record; this WP is useful for tracing the conceptual history.