Do public insurance programs crowd out private savings? I examine the relationship between Medicaid and wealth and make a contribution to the literature on this issue in three primary ways. First, I apply the instrumental-variables approach developed by Gruber and Yelowitz (1999) to a different dataset, the National Longitudinal Survey of Youth, 1979 (NLSY79), while at the same time examining an alternative instrument. The results turn out to differ depending on the instrument and, for one of the instruments, to be sensitive to assumptions needed to identify Medicaid’s effects. Second, using the longitudinal data in the NLSY79, I am able to observe families before and after becoming eligible for Medicaid, and use fixed-effects to control for family-specific unobservable factors that are correlated with both Medicaid eligibility and wealth accumulation. It turns out, however, that assessment of the impact of Medicaid by means of fixed effects has its limitations as well. Third, I make use of the SIPP data used by Gruber and Yelowitz themselves, and examine the sensitivity of their conclusions to omitted factors that may be related to both Medicaid eligibility and to wealth accumulation. While more robust than the results using the NLSY79, the SIPP estimates are found to depend on the sample used and on certain specification restrictions. Taken together, the results suggest caution in making inferences about the impact of Medicaid on wealth.