Estimating the consumption-smoothing value of Medicaid expansion: Evidence and limits

We ask what direct consumption data can reveal about the consumption-smoothing component of the insurance value of the Affordable Care Act Medicaid expansions. Medicaid expansion may reduce consumption risk by limiting exposure to out-of-pocket medical spending, but this channel may be muted when much uninsured care is financed through charity care, bad debt, or other forms of uncompensated care. We combine household expenditure data from the Consumer Expenditure Survey with the staggered timing of state Medicaid expansions to estimate effects on average consumption and on the within-education-group consumption distribution. We then map these estimated distributional effects into a static expected-utility framework. The exercise highlights both what this approach can reveal and where it is limited. Point estimates are largest for individuals without a high school degree and close to zero for higher-education groups, but the estimates are imprecise, including for average consumption. The resulting insurance-value calculations require additional assumptions and are also statistically imprecise. We therefore interpret the results as showing that existing consumption data can be used to bound the consumption-smoothing channel, but are not adequate to precisely estimate this component of the insurance value of Medicaid expansion. The findings clarify what can and cannot be learned from consumption data about this component of Medicaid’s value.