Health plays an important role in the labor market. Human capital accumulation, occupational choice, hours, wages, turnover, and retirement are all both causes and consequences of health. Better health leads to higher productivity, and healthy individuals can spend more time on the labor market. Yet, at the same time, health itself is also a function of past labor outcomes. High earners may have access to better health care and can afford better nutrition and housing. However, more income and working hours often come at the expense of psychological stress, which in turn can decrease future performance measures. This complex interdependency between health and labor makes it difficult to isolate causal mechanisms. Nevertheless, understanding them is crucial to design cost-effective policies that target workforce health. In this thesis I present three papers which use causal identification strategies and big administrative datasets to shed light on some of the mechanisms briefly discussed above.