Homelessness—and its precursor, housing instability—is associated with adverse health. However, inferring causal pathways is challenging because people who suffer unstable housing often experience many other disadvantages. This paper consists of two parts. The first part uses the staggered rollout of New York City’s Universal Access to Counsel (UA) program as a natural experiment to study the impact of housing court involvement on health and homelessness. The second part focuses on homelessness. We assemble comprehensive descriptive evidence about shelter entrants’ preexisting relationships with medical providers and patterns of health care utilization. We then analyze how shelter entry affects health and health care trajectories, both in general and as a function of certain attributes of shelter assignments, such as the proximity of shelters to pre-shelter primary care providers. We are especially interested in the hypothesis that the housing stability and support services afforded by temporary shelter—as well as, in some cases, ensuing permanent housing placements—may improve health in underappreciated ways.