The growth of novel flexible work formats raises a number of questions about their effects upon health and the potential public policy implications. However, answering these questions is hampered by data and identification constraints. This is the first paper that draws on comprehensive longitudinal administrative data to examine the impact of self-employment in terms of health. In addition to variation in work status of each individual over time, we also consider variation driven by a number of novel instrumental variables. We also focus on an objective health outcomeospital admissionsthat is not subject to recall or other biases that may affect previous studies. Our findings, based on a representative sample of over 100,000 individuals followed monthly from 2005 to 2011 in Portugal, indicate that self-employment tends to reduce the likelihood of hospital admission by at least half.