I exploit the seasonality of the Earned Income Tax Credit, a large transfer to low-income working families, in order to estimate the effect of cash liquidity on the demand for medical care for low-income adults in the United States. I use the estimate to decompose moral hazard into a welfare increasing ``liquidity effect'' and a welfare decreasing ``substitution effect.'' Empirical estimates of the liquidity effect are large. In months when the EITC arrives, individuals are more likely to use any medical care. Visits to the doctor's office, outpatient hospital treatment, and prescription drugs purchase increase, but there is little evidence of an increase in the use of emergency room, inpatient, or dental care. I then use the new empirical estimate to evaluate the net change to social welfare resulting from an expansion of public health insurance like the Affordable Care Act. Estimates suggest that, at the margin, the benefit of increasing public insurance benefits for low-income adults exceeds the cost.