Using a Törnqvist chained-index technique applied to Social Security Administration (SSA) administrative data (1 percent Continuous Work History Sample [CWHS]), Pattison and Waldron decompose the 253% growth in new disabled-worker entitlements from 1970 to 2008 into three components: overall working-age population growth, growth in the proportion of the population insured for disability (driven by rising female labor force participation), and the residual disability incidence rate — each further split by age/sex adjustment. The headline result: 90% of 36-year growth is attributable to demographic factors after age/sex adjustment; the adjusted incidence rate explains only 10%. In the 1990–2008 subperiod, demographics explain 94% and adjusted incidence only 6%. The paper's most important methodological finding is that the unadjusted incidence rate creates an apparent post-1990 acceleration that disappears entirely once the baby boom aging effect is controlled for.
| Period | Insured workers (incl. pop + insured-rate) | Adjusted incidence rate |
|---|---|---|
| 1972–1990 | 83% | 17% |
| 1990–2008 | 94% | 6% |
| 1972–2008 | 90% | 10% |
Unadjusted figures (for comparison, showing the baby boom illusion):
| Period | Insured workers | Unadjusted incidence rate |
|---|---|---|
| 1972–1990 | 139% | −39% |
| 1990–2008 | 43% | 57% |
| 1972–2008 | 81% | 19% |
The 139%/−39% split in 1972–1990 reads as: insured workers grew faster than entitlements (partly because early baby boom growth was at young, low-incidence ages). The apparent 57% incidence share in 1990–2008 evaporates once the baby boom is controlled for.
"Although the raw or unadjusted growth in the number of workers becoming entitled to benefits under Social Security's DI program gives the appearance of an upward and accelerating trend, using such a measure may lead to misleading analytical conclusions."
"The adjusted incidence rate can be considered the residual growth that is unexplained after taking into account all the easily observable factors: population growth, the increase in the proportion of the population that is disability insured…and the interaction of the shifting age composition of the insured population with the age pattern of disability incidence."
The unadjusted/adjusted inversion is the paper's central result and is genuinely striking: the same data that appear to show a post-1990 incidence explosion instead show incidence slowing once the baby boom is stripped out. This directly undermines any narrative that treats rising post-1990 DI caseloads as evidence of genuine deterioration in population health or loosening eligibility standards. The incidence/prevalence distinction is also under-appreciated: Liebman (2015) and Autor-Duggan analyze the beneficiary ratio, which conflates the entry rate with mortality, recovery, and Old-Age and Survivors Insurance (OASI) conversion trends; Pattison/Waldron isolate just the entry rate, which is arguably what policy most directly controls. The 10% residual finding is important for calibrating concern: even if rising incidence were entirely a policy failure, it explains only a tenth of 36-year growth. The paper is methodologically rigorous but deliberately agnostic about what drives the residual — disentangling health, economic, and policy contributions within that 10% is left to other work.