Lindner Burdick and Meseguer 2017 — Characteristics and Employment of Applicants for Social Security Disability Insurance over the Business Cycle

Summary

Using the universe of all adult disability insurance (DI) applications 1991–2008 (22.7 million, Social Security Administration (SSA) Disability Research File) in a state-level panel regression, the paper examines how the composition of DI applicants changes with unemployment. A 1 percentage point (pp) rise in unemployment increases applications by ~3.1% and shifts composition toward conditional applicants — those with higher work capacity who prefer employment and apply only after job loss. The paper then shows that post-application outcomes of denied applicants are slightly negatively related to unemployment, because the labor market suppression effect outweighs the compositional improvement. A calibrated model shows the disincentive effect of the program declines with unemployment.

Key Claims

• A 1 pp rise in unemployment → ~3.1% increase in DI applications. Step 2 and Step 4 applicant shares rise; Step 3 (Listing-eligible; most severe) is recession-invariant in both number and share.
• Pre-application earnings of denied applicants rise with unemployment (7–5 years before application: ~$200 higher per 1 pp rise) — consistent with a shift toward conditional applicants with stronger work histories.
• Post-application earnings and employment of denied applicants are negatively related to unemployment: earnings ~$305 lower 2 years after application; employment ~1 pp lower 2 years after; both fade by year 5.
• Application processing time is ~2–3 days shorter per 1 pp unemployment rise; acceptance rates fall at all stages (initial, final, and including re-applications).
• Of all applicants, 40.6% are accepted at initial determination, 57.5% at final determination, and 65.5% including re-applications within 5 years — a large gap driven by the appeals process.
• The disincentive effect (counterfactual earnings of accepted applicants had they not enrolled) is slightly negatively related to unemployment: the recession suppresses employment prospects for the whole low-skill workforce, not just DI applicants, making DI less distortionary precisely when most heavily used.
• The share of joint DI/Supplemental Security Income (SSI) applications rises with unemployment, despite higher-work-capacity applicants presumably having higher income — explained by income losses around the time of application.

Concepts Introduced or Extended

• Conditional DI Applicants
• SSA Sequential Determination Process
• DI Denied Population
• Disability Insurance Chronology
• Disability Insurance Program

Entities Mentioned

• Stephan Lindner
• Clark Burdick
• Javier Meseguer
• Social Security Administration

Quotes

"Our findings caution against equating increases in the share of conditional applicants and beneficiaries with an increase in the disincentive effect of the program. Rather, it is essential to identify the cause of these compositional changes."

"Instead of going through the lengthy appeals process that might further diminish their chances of finding new employment, it might be better to provide short-term financial support together with re-employment services to them."

My Take

The paper's cleanest contribution is the step-specific decomposition: Step 3 is recession-invariant while Steps 2 and 4 surge — this is a direct empirical fingerprint of the conditional applicant mechanism rather than just a general increase in need. The disincentive effect finding is the most policy-relevant result: it undercuts the recession-era narrative that DI is wasting resources on people who could work. The appeal-trap policy implication is important but acknowledges the countervailing difficulty: even less-severely-disabled recession applicants face real employment barriers. The paper covers 1991–2008 and explicitly sets aside the Great Recession as a robustness check; extending the analysis post-2008 is a natural next step.