There are two health insurance systems in the United States for working-age individuals: (1) employer-sponsored coverage for employed individuals and their families; and (2) public health insurance (Medicaid and Medicare) for individuals who are deemed unable to work.  Tying health insurance to employment has well-known adverse side-effects, often referred to as “job lock” (Gruber 2000, Gruber and Madrian 2004).  Research suggests that job lock decreases job turnover, decrease entrepreneurial activity, and influences retirement timing (Gruber and Madrian 1994, Madrian 1994, Buchmueller and Valletta 1996).  Similar adverse side effects could apply to welfare recipients or the disabled population; tying health insurance coverage to receipt of cash benefits may exacerbate the already strong incentives to stay on the welfare/disability rolls.  Evidence suggests that “welfare lock” is statistically significant, though relatively small in magnitude (Ellwood and Adams 1990, Yelowitz 1995).  The importance of DI lock remains unquantified, despite the substantial value of public health insurance benefits to the disabled (Autor and Duggan 2006). 

There is considerable policy concern about “DI lock” – that tying public health insurance coverage to cash disability benefit receipt contributes to the low exit rates due to work.  This concern led Congress to institute continued health insurance eligibility after disability beneficiaries leave the cash-benefit rolls for work-related reasons.  However, unlike the long literature on “job lock,” the importance of the DI lock hypothesis – either before or after these extensions – has remained unquantified. 

 This paper tests whether a “perceived DI lock” remains among disability beneficiaries, and whether state health insurance policies help alleviate the problem and encourage work among beneficiaries.  The analysis includes both DI and SSI beneficiaries, and tests if there are differential patterns between the two programs.  We exploit state variation in the access and cost of health insurance, caused by regulation of the non-group market, the existence of Medicaid buy-in programs, and Medicaid generosity. 

Overall we find some evidence of persistent DI-lock which Medicaid buy-in programs help alleviate.  Heterogeneity is very important in this context.  Our estimates suggest that increasing health insurance access substantially increases the likelihood of positive earnings among a subset of disability beneficiaries.  We find evidence of SSI lock among beneficiaries with some Medicaid expenditures, and that both non-group health insurance regulation and generous Medicaid eligibility help alleviate the problem.  We find evidence of remaining DI lock among individuals who do not have access to supplemental health insurance outside of Medicare.  Medicaid buy-in programs alleviate the remaining DI lock.