Using data from the Moving to Opportunity randomized housing voucher experiment, we estimate the direct effects of housing and neighborhood quality on child health. We show that, five years after random assignment, housing mobility has little impact on overall health status, asthma, injuries, and body mass index. The few effects that we observe imply that being offered a voucher through the program might worsen some aspects of child health, despite significant improvements in housing quality, nutrition and exercise, and neighborhood safety. Our results are inconsistent with the hypothesis that neighborhood conditions explain much of the widely-cited income gradient in child health. (Contains 22 footnotes, 7 tables, and 1 figure.)