Ruffing 2015 — Geographic Pattern of Disability Receipt Largely Reflects Economic and Demographic Factors

Summary

A 7-page descriptive policy brief from the Center on Budget and Policy Priorities (CBPP) showing that four demographic and economic factors explain approximately 85% of cross-state variation in Disability Insurance/Supplemental Security Income (DI/SSI) receipt rates among the working-age population (ages 18–64, 2013 data): low educational attainment, older workforce, fewer immigrants, and blue-collar industry mix. The paper reframes the "geography of disability" — South and Appalachia rates nearly double the national average — as a structural consequence of the Social Security Administration's (SSA) own eligibility criteria operating on a labor force that is less educated, older, and more physically employed than the national average, not as evidence of program abuse or lax screening.

Key Claims

• About 6% of the working-age population (ages 18–64) receives DI or SSI nationally. States with the highest combined rates: West Virginia (12.5%), Kentucky (11.4%), Alabama (11.0%). States with the lowest: Utah (3.8%), Alaska (4.1%), Hawaii (4.1%).
• Education is the single strongest predictor: states with low high-school completion rates have high disability receipt, directly reflecting the SSA vocational eligibility framework, which requires less-educated severely impaired workers to show they cannot realistically retrain for other, less demanding work.
• Unemployment rates explain almost none of the geographic variation. Although higher unemployment boosts applications, researchers find those marginal applicants are also more likely to be denied — the programs' eligibility rules do not become more lax in periods or areas of high unemployment. This is consistent with Lindner, Burdick, and Meseguer (2017).
• Immigrant-dense states (California, New York, New Jersey, Florida, Nevada, Texas) have lower rates due to DI work-history requirements (typically at least five of the last ten years in covered employment) and the 1996 SSI ban on benefits to new immigrants.
• Poverty rate explains SSI participation (means-tested) but has little bearing on DI participation (not means-tested).
• Within-state variation can be as large as between-state variation: Virginia simultaneously had 5 of the top-10 highest-rate counties and 4 of the top-10 lowest-rate counties nationally in 2011, reflecting intrastate heterogeneity in age, education, and industry mix.
• Four factors combined explain ~85% of cross-state variation (ordinary least squares [OLS] regression). The remaining variation reflects hard-to-measure factors including public awareness, legal assistance availability, historical (not current) industry mix, and Affordable Care Act (ACA) insurance expansion effects.

Concepts Introduced or Extended

• Vocational Grid — the education factor is the geographic expression of the vocational grid mechanism: less-educated workers with serious conditions have far fewer retraining options under the "any work in the national economy" standard
• Conditional DI Applicants — the blue-collar industry and education factors identify the structural pool of conditional applicants at the geographic level
• Disability Insurance Program
• SSI Children's Program

Entities Mentioned

• Kathy A. Ruffing
• Center on Budget and Policy Priorities
• Social Security Administration
• Nicole Maestas
• Kathleen Mullen
• Alexander Strand

Quotes

"Those four factors alone are associated with about 85 percent of the variation in disability receipt rates across states."

"A surgeon who suffers a stroke might become a pathologist or an insurance consultant, but that's hardly feasible for a 60-year-old waitress without a high-school diploma who has also suffered a stroke."

"Unemployment rates explain hardly any of the geographic variation. Although higher unemployment boosts disability applications, researchers conclude that those extra applicants are also more likely to be denied. The programs' eligibility rules are strict and don't become more lax in periods or areas of high unemployment."

My Take

The 85% $R^2$ claim comes from an OLS regression with four cross-state predictors — the causal logic is well-motivated by program rules but the analysis is purely correlational. The paper's framing is explicitly defensive of the program against claims that geographic variation reflects lax screening; it is advocacy analysis, not econometrics. The most analytically solid point is that unemployment barely explains geographic variation, consistent with the research literature showing that cyclical application surges don't translate into award surges. The surgeon-vs.-waitress example is an effective illustration of why SSA's vocational criteria mechanically produce the observed geographic distribution: less-educated workers with serious impairments have far fewer retraining options under the "any work in the national economy" standard, so their claims are more likely to be allowed at Step 5.