I estimate the causal impact of Medicaid eligibility on take up, private health insurance coverage, healthcare utilization, and children's health by using a regression discontinuity design. In contrast to a standard regression discontinuity design, identification exploits multiple thresholds that arise from variation across states in income eligibility rules. Using data from the Panel Study of Income Dynamics and its Child Development Study supplement, I find that Medicaid eligibility increases take up by 10–13 percentage points on average, rising to 24–29 percentage points at lower income eligibility thresholds. There are significant crowding out effects of the same magnitude as those on take up rates. Medicaid eligibility increases the use of preventive health care by 11–14 percentage points but only at low income thresholds. Finally, I find that Medicaid eligibility has no significant effects on health outcomes in the short and medium run. Copyright © 2012 John Wiley & Sons, Ltd.