The Relation Between Health and Earnings in Self-Employment.

Multiple studies have shown that, on average, the self-employed are healthier than wage workers. The link between the health of self-employed individuals and their financial performance in terms of earnings is, however, less understood. Based on human capital theory, we expect a positive link between health and earnings among the self-employed. For two reasons we expect the relationship between health and earnings to be stronger for the self-employed than for wage workers. First, the self-employed can more easily adapt their production activities such that they yield the highest returns to their human capital, including their health. Second, in the short term, the earnings of the self-employed are more dependent on the ability to work than the wages of wage workers. Our empirical analysis draws on data from the Household, Income and Labor Dynamics in Australia (HILDA) survey, a longitudinal dataset (2001-2017). Our outcome variable is an individual's total income derived from wage work and/or running a business. Health is measured using multi-item constructs for , , and from the Short Form Health Survey (SF-36). We distinguish between wage workers and self-employed individuals with and without employees. Fixed-effects regressions reveal a significant positive relationship between health and earnings in self-employment as well as in wage work. As expected, this relationship is significantly stronger in self-employment than in wage work (for and , but not for ). The latter result holds particularly for self-employment without employees. We provide evidence that the higher returns can be partly explained by the fact that the earnings in self-employment are more dependent on the ability to work (as proxied by the number of working hours) than earnings in wage work. We also find a negative relationship between health and job termination. Again, this relationship is stronger for the self-employed (without employees) than for wage workers (for and , but not for ).