I test whether control beliefs buffer labor-supply responses to episodes of ill health using German panel data. Men with negative versus men with positive control beliefs are 100% more likely to drop out of the labor market a year after the health shock. A series of robustness checks shuts off alternative explanations for heterogeneous responses. The buffering effects are most beneficial to men from disadvantaged backgrounds. Long-term labor market and health benefits of control beliefs are likely to be large.