OLeary et al 2012 — Workplace Injuries and the Take-Up of Social Security Disability Benefits

Summary

Using matched New Mexico (NM) Workers' Compensation (WC) Administration records (injury dates 1994–2000), Social Security Administration (SSA) administrative data, and Internal Revenue Service (IRS) earnings files, the authors estimate the causal impact of workplace injuries on subsequent Social Security Disability Insurance (DI) receipt. Medical-only cases (≤7 days off work) serve as the counterfactual for uninjured workers; lost-time cases are the treatment. The core finding: a lost-time workplace injury roughly doubles the 10-year probability of DI receipt (6% → 12%), and the effect is consistent across all age groups — equating in magnitude to a 10-year aging effect. The paper also quantifies fiscal costs and documents DI's role as backup insurance when workers' compensation replacement rates fall below policy targets.

Key Claims

• N = 98,148 cases (65,705 medical-only, 32,443 lost-time); injury dates 1994–2000; followed through December 2009.
• Lost-time injury doubles 10-year DI receipt probability: 6% medical-only → 12% lost-time (Kaplan-Meier, Chart 2).
• "10-year aging effect": within each age group, lost-time cases match the DI hazard of medical-only cases in the next older 10-year cohort (Chart 3). Applies across all groups (15–24, 25–34, 35–44, 45–54).
• 15-year cumulative DI receipt by group: ages 15–24: 3.0%/4.9%; ages 25–34: 5.5%/10.1%; ages 35–44: 10.9%/20.0%; ages 45–54: 20.3%/34.4% (medical-only/lost-time).
• Severity gradient: permanent total disability (PTD) > permanent partial disability (PPD) > temporary disability (TD) ≥8wk > TD <8wk > medical-only. Even TD <8wk cases show substantially higher DI risk than medical-only.
• Cox proportional hazards (stratified by 10-year age group due to proportionality violation): preinjury earnings <$10K → hazard ratio (HR) = 2.40 (95% confidence interval [CI]: 2.11–2.73) vs. $50K+ reference; earnings gradient monotonic and large.
• Mining industry: highest industry hazard ratio (medical-only: 1.42; lost-time: 1.27); construction and health also elevated for medical-only.
• Women: lower DI hazard (HR ≈ 0.82 medical-only, 0.90 lost-time vs. men).
• WC-attributable DI: ~50% of new DI awards among injured NM workers; ~7% of all NM DI awards; extrapolated nationally → ~70,000 new DI awardees in 2010 attributable to WC injuries.
• Fiscal extrapolation: first-year SSA cost ~$889M; present value ~$12B per annual cohort; doubles with Medicare.
• DI as backup insurance: WC replaces well under half of long-term lost earnings in PPD/long-term TD cases (Boden and Galizzi 1999; Reville 1999); DI absorbs the remainder for severely impaired workers.
• WC-DI offset: combined benefits capped at 80% of preinjury earnings; ~11% of DI beneficiaries receive WC/public disability benefits (PDB); their DI is ~6% lower due to offset (Newsome and Parent 2008).
• WC-DI substitution at aggregate level: WC and DI trends inverse 1980–2007 (Sengupta, Reno, and Burton 2011); but McInerney and Simon (2012) find no within-state substitution — the aggregate inverse pattern is driven by between-state variation, not individual-level switching.
• Employer incentive failure: employers do not bear full injury costs → suboptimal prevention incentives → DI effectively subsidizes safety failures (payroll-tax-funded, not risk-adjusted).
• Delayed injury effects: disability determination typically occurs 1–2 years postinjury; DI enrollment often years later as cumulative health, labor market, and functional effects compound.
• WC cash benefits grew +65% (1987–2008) vs. DI +403% — divergent growth despite serving the same population.

Concepts Introduced or Extended

• Workers' Compensation and DI
• Conditional DI Applicants
• DI Application Costs and Take-Up
• Disability Insurance Program

Entities Mentioned

• Paul O'Leary
• Leslie Boden
• Seth Seabury
• Social Security Administration

Quotes

"Currently, however, workers' compensation often provides replacement levels that fall short of its own ideals. In those cases, DI potentially acts as backup insurance, reducing the financial burden of the long-term consequences of workplace injuries on the most severely disabled." (p. 13)

"Employers, who are in the best position to improve workplace safety, do not bear the full costs of those injuries and therefore have a reduced incentive for prevention. Instead, employers shift some of the costs to workers and to the DI program, which workers and employers fund through payroll taxes that are not risk adjusted." (p. 13)

My Take

This is the most compelling empirical case for the occupational injury pathway into DI — a mechanism that the standard US literature (which uses admin data without WC linkage) largely ignores. The "10-year aging" framing is unusually vivid and policy-relevant: it reframes the DI cost question in terms of injury burden rather than eligibility laxity. The medical-only-as-control identification is reasonable but not fully convincing (medical-only workers may still have systematic health differences from truly uninjured workers). The single-state design is a real limitation; NM's economy (mining-heavy, lower wages) may not generalize. The fiscal extrapolation is illustrative rather than causal. The most important conceptual contribution is the documentation that WC and DI are not substitutes at the individual level (contra the aggregate inverse trend), and the backup-insurance framing that has direct implications for employer liability reform.