Association for Public Policy Analysis & Management

Research Objective: To estimate the effect of Medicaid coverage on self-reported measures of access to care, affordability of care, utilization, and health status.

Study Design: We use data from the 2009-2014 National Health Interview. We exploit policy variation created by state decisions to opt out of the ACA Medicaid expansion and estimate instrumental variables models that isolate the effects of Medicaid coverage for non-disabled adults below 100% of the federal poverty level who gained coverage because of state decisions to adopt the expansion. Because Medicaid eligibility prior to the ACA varied widely among states adopting the expansion, we separately analyze the impacts of the 2014 expansion in two groups of expansion states: those where some or all adults in poverty first gained Medicaid eligibility in 2014 vs. states where all adults in poverty were eligible for Medicaid in 2013 and earlier years. One might expect Medicaid coverage to have different impacts in these states due to differences in the composition of the population taking up coverage in response to policies implemented in 2014.

Principal Findings: We find that Medicaid coverage led to significant improvements in the self-reported affordability of care in states that expanded eligibility in 2014, but not in states where adults in poverty were previously eligible. In the 2014 expansion states, Medicaid coverage substantially reduced the likelihood that individuals had delayed care due to cost in the past year, went without care due to cost in the past year, or reported being very worried about paying for care if they became sick or injured. On other measures of access to care and affordability, however, we failed to find significant evidence that Medicaid coverage led to substantial changes in individuals' usual place of care, the probability that individuals reported having trouble finding a doctor or affording medication, or the probability that individuals were currently paying off a medical bill over time. We also found no significant evidence that Medicaid coverage led to changes in self-reported health status or mental health for individuals gaining Medicaid coverage in 2014 in either group of states.

A likely explanation for the differences in impacts across the two groups of states is that there were different patterns of changes in insurance from 2013-2014 across these groups of states. We find a similar increase in Medicaid coverage across both groups of states, but the net increase in insurance was only about half as large in the states where these adults were previously eligible because those states saw significant reductions in private insurance coverage (i.e., crowd-out). Even so, adults in both groups of expansion states were much more likely to report that their health insurance was better than a year ago.

These findings suggest that differences across states in pre-ACA Medicaid policies are likely to lead to meaningful differences in the impact of Medicaid expansion on insurance status. Future evaluations of the ACA Medicaid expansion may benefit from separate analyses of insurance choices and access to care for the newly eligible and previously eligible populations.