Duggan, Singleton, and Song 2007 — Aching to Retire The Rise in the Full Retirement Age and Its Impact on the Disability Rolls

Summary

Estimates the behavioral effect of the 1983 Social Security Reform Act's Full Retirement Age (FRA) increase on disability insurance (DI) enrollment. The FRA increase reduced the generosity of early retirement benefits at age 62 (from 80% to 70% of the Primary Insurance Amount (PIA) by 1960 cohorts) without any corresponding change to DI benefits, widening the financial advantage of DI over early Old-Age (OA) benefits and increasing the incentive to apply. Using Social Security Administration (SSA) Continuous Work History Sample data for men born 1935–1945 (715,968 person-year observations), the paper identifies a statistically significant but quantitatively small effect concentrated among men ages 63–64 and low-income workers. Long-run estimate: ~42,000 additional male DI recipients (1.3% of male rolls), offsetting less than 4% of projected Old-Age, Survivors, and Disability Insurance (OASDI) savings from the FRA increase.

Key Claims

Rising DI incentive: Early retirement benefit at 62 falls from 80% of PIA (born ≤1937) to 75% (born 1943–1945) to 70% (born 1960+); DI unchanged. Expected present value (PV) advantage of DI over early OA rises from ~$36,793 (born 1937) to ~$45,991 (born 1943) to ~$55,190 (born 1960)
Age concentration: Effect is statistically significant and concentrated almost entirely among men ages 63–64 — the cohort-year just past early retirement. The estimated effect for 63–64-year-olds is ~5× larger than for 56–60-year-olds
Income concentration: Effect significant for men below median Average Indexed Monthly Earnings (AIME), null for those above — consistent with lower health, lower opportunity cost of applying, and higher income share from Social Security for low earners. Low-income workers in the lowest AIME decile face mortality rates >3× those in the top decile, reducing the PV gain and partly offsetting the incentive for high earners
Long-run magnitude: ~42,000 additional male DI recipients once fully implemented (FRA = 67); 1.6 percentage-point (pp) increase in DI enrollment for 63–64-year-old men (10.0% → 11.6%); just 1.3% of total male DI enrollment
Fiscal offset: The induced DI increase offsets only 3.9% of projected OASDI savings from the FRA rise — the reform remains fiscally sound despite the DI margin
Placebo test: False assignment of the policy to 1927–1937 cohorts (unaffected by 1983 legislation) produces null results in all specifications, validating the causal interpretation

Concepts Introduced or Extended

Entities Mentioned

Quotes

"While it is clearly difficult to estimate the eventual effect of the policy change given that it is not yet fully implemented and given that many other factors are likely to change over time, our estimates suggest an additional 42,000 male DI recipients in the long-run. This forecast holds fixed the age distribution of men insured for DI benefits at 2003 levels and would represent an increase of just 1.3 percent in the number of men receiving DI benefits."

My Take

A clean identification paper: year-of-birth variation in the OA/DI ratio is genuinely exogenous, the mechanism is transparent, the placebo passes cleanly, and the result is precisely estimated. The ~1.3% magnitude is the paper's most policy-relevant finding — the FRA behavioral substitution channel exists but is small enough that it does not seriously undermine the fiscal case for raising the FRA. This is distinct from Ruffing's (2014) mechanical FRA effect (+450K who remain on DI rolls until the new FRA rather than converting at 65); together the two channels quantify both pathways through which the 1983 FRA reform expands DI rolls. The exclusion of women (due to rising female labor force participation (LFP) changing sample composition over cohorts) and younger workers likely understates the true total effect.