Medicare Eligibility and Labor Force Exit:\ Effects by Automation Exposure

Abstract

How does Medicare eligibility at age 65 affect labor force participation, and does this effect differ by workers' exposure to automation risk? Using Current Population Survey data from 2015–2024 and a regression discontinuity design at the Medicare eligibility threshold, we find that Medicare eligibility reduces labor force participation by 2.8 percentage points overall using local linear regression with robust bias-corrected inference. Critically, this effect is larger for workers in education groups associated with higher automation exposure: workers with high school education or less experience a 3.6 percentage point decline (95% CI: $[-7.2, -1.6]$), compared to 2.2 percentage points for college-educated workers (95% CI: $[-4.6, 0.2]$). Covariate balance tests confirm no discontinuities in observable characteristics at the threshold. These findings suggest that Medicare eligibility may release "job lock" more strongly for workers in occupations facing greater automation risk, potentially because these workers have less bargaining power with employers and fewer labor market alternatives. The results have implications for understanding how the interaction of health insurance policy and technological change shapes retirement decisions.