Summary
Livermore and Bardos (2014) document that approximately 28% of Social Security Disability Insurance (SSDI)-only beneficiaries (those not concurrently receiving Supplemental Security Income, SSI) live in households below the federal poverty line, despite having work-history-based benefits. Using four pooled waves of the National Beneficiary Survey (NBS) (2004, 2005, 2006, 2010; n = 6,045), they profile the personal, health, employment, and income characteristics of poor versus higher-income SSDI-only beneficiaries. Marital status is the strongest single predictor of poverty: unmarried beneficiaries are 23 percentage points (pp) more likely to be poor than married beneficiaries.
Key Claims
- 28% poverty rate: Nearly one-in-three SSDI-only beneficiaries lives below 100% of the federal poverty level, despite receiving work-history-based disability insurance benefits averaging $1,146/month (December 2013).
- Marriage dominates: Unmarried SSDI-only beneficiaries are 23 pp more likely to be in poverty (predicted probability 37% vs. 14%). This is the strongest predictor in the logit model, stronger than education, health, or employment status.
- Education second: Less than high school (<HS) education raises poverty probability by 12 pp (36% vs. 24%).
- Employment rates equal, aspirations diverge: Poor and higher-income beneficiaries are employed at statistically indistinguishable rates (~8–10% at interview). But poor beneficiaries are significantly more likely to have employment goals (41% vs. 30%) and to have recently sought work. This gap between aspiration and reality implicates structural barriers, not motivation.
- Structural barriers: Poor non-working beneficiaries more often cite transportation, inability to find qualifying jobs, employer discrimination, fear of losing cash/health benefits, and need for additional training as barriers to work.
- Lower SSDI benefits: Poor beneficiaries receive lower average SSDI benefits, reflecting weaker pre-disability earnings histories — consistent with earlier disability onset, less education, and more interrupted work careers.
- Public assistance reliance: Poor beneficiaries are more likely to have Medicaid (50% vs. 11%) and other means-tested support. Most (55%) receive <$1,000/month in total government assistance.
- Childhood disability onset: 18% of poor beneficiaries experienced disability onset before age 18 vs. fewer in higher-income groups. However, childhood onset is not a significant poverty predictor after controlling for education, marital status, and other characteristics — suggesting it operates through those channels.
- Diagnostic composition: Poor beneficiaries are ~2× more likely to report psychiatric conditions (20% vs. 11%) or intellectual disability (6% vs. 3%) as their primary limitation. Sensory disability is associated with significantly lower poverty probability.
- Sheltered employment: Among the employed, poor beneficiaries are substantially more likely to be in sheltered or supported work settings (42% vs. 24%).
Concepts Introduced or Extended
Entities Mentioned
Quotes
"For most SSDI beneficiaries, disability benefits make up a critical source of support, representing 75 percent or more of beneficiaries' monthly income."
"Although employed at levels similar to higher-income beneficiaries, poor beneficiaries were significantly more likely to have employment goals and expectations. This, along with other differences between the two groups, suggests that poor beneficiaries likely face bigger barriers to employment than beneficiaries with higher income."
My Take
A useful descriptive brief but methodologically limited. The NBS pooled sample gives adequate power for the descriptive comparisons; the logit model identifies correlates of poverty, not causes. The paper's main contribution is the clean documentation of the aspiration-employment gap among poor SSDI beneficiaries — this is a meaningful finding because it refutes a narrative that poor beneficiaries are poverty-trapped by low work motivation or benefit-cliff disincentives alone. The structural barriers they report (transportation, hiring discrimination, benefit cliffs) are more likely culprits, but the brief does not have the causal identification to establish this.
The "gap population" framing — too rich for SSI, too poor to escape poverty — is policy-relevant and underexplored in the literature. The brief does not estimate how large the welfare gain would be from closing this gap, nor does it examine whether the gap is widening or narrowing over time. The 2010 data inclusion (post-recession) slightly complicates the employment numbers but the authors argue it did not meaningfully affect poverty classification rates.