Moore 2015 — The Employment Effects of Terminating Disability Benefits

Summary

Moore (2015) exploits the January 1997 termination of 65,000 DA&A (drug addict and alcoholic) disability insurance (DI)/Supplemental Security Income (SSI) beneficiaries — a consequence of the 1996 Contract With America Advancement Act removing drug and alcohol addiction as qualifying conditions — as a natural experiment to estimate the employment effects of losing DI benefits. Using Social Security Administration (SSA) earnings records, he finds a ~22 percentage point (pp) rise in employment above substantial gainful activity (SGA) peaking in 1998–2000, decaying to ~9pp by 2008 primarily through re-qualification on the rolls rather than health deterioration. The most policy-actionable finding is an inverted-U relationship between DI duration and employment response: beneficiaries terminated after approximately 2.7 years on the rolls show the largest effects (24.6pp), substantially larger than both shorter- and longer-duration cohorts.

Key Claims

• The 1997 DA&A termination produced a ~22pp peak employment response (above SGA) among ~65,000 terminated beneficiaries — large relative to the <1% annual voluntary exit rate among DI recipients.
• Employment response is an inverted-U function of DI duration: peaks at ~2.7 years of receipt (24.6pp), exceeds both the 9-month cohort (16.3pp) and the 6-year cohort (18.8pp).
• By 2008, employment effects decay to ~9pp. The primary driver is re-qualification: terminated beneficiaries returned to DI through re-application, not because their employment ended. The employment "fade" substantially overstates any genuine return to disability.
• Initially-awarded beneficiaries (deemed most severely disabled at initial application — they were granted without a hearing) match hearings-awarded beneficiaries in employment response after 5–6 years of DI receipt, consistent with health improvement or mean reversion while on the rolls.
• Heterogeneity: younger workers respond more (ages 30–39: ~25pp vs. 50–61: ~16pp); higher pre-DI earnings → larger response; lower local unemployment → larger response (labor demand constraints bind).
• External validity caveat: DA&A beneficiaries are a specific subgroup who may have more residual work capacity than the average DI recipient. Moore acknowledges this limitation explicitly.

Concepts Introduced or Extended

• Causal Effects of DI Receipt
• Disability Insurance Program — continuing disability review (CDR) optimal timing (2–3 years); time-limited benefit design
• Substantial Gainful Activity — used throughout as the employment outcome measure; validated as a detectable and meaningful threshold

Entities Mentioned

• Timothy Moore
• Social Security Administration

Quotes

"The results suggest that policy interventions such as Continuing Disability Reviews will be most effective if targeted at beneficiaries who have been receiving benefits for around two to three years."

"At the peak employment response, individuals who had been receiving benefits for around 2.7 years were 24.6 percentage points more likely to be employed above SGA, which is about 50 percent higher than the response for those in their first year of benefits."

My Take

The inverted-U duration finding is the paper's most important result: it refutes the intuition that "longer on DI → harder to return to work monotonically," and instead identifies an optimal intervention window around 2–3 years. The implication for CDR scheduling is specific and actionable. The external validity limitation (DA&A subgroup) is real but partially mitigated by consistent heterogeneity patterns across age and earnings groups. The re-qualification pathway — terminated beneficiaries returning to DI by 2008 — complicates welfare conclusions: it is consistent with both "the termination was appropriate but they eventually re-qualified legitimately" and "many were genuinely still disabled and the termination was harmful." Moore does not attempt to adjudicate between these interpretations, which remains an open policy question.