Occupational Choice in Rural Kenya: Using Subjective Expectations Data to Measure Credit and Insurance Constraints∗

Millions of people living in rural areas of Sub-Saharan Africa supplement earnings from subsistence farming by operating microenterprises or by working as day laborers. However, little is known about how people choose which of these additional incomegenerating activities to do, and, in particular, to what extent market imperfections affect this choice and contribute to inefficiencies in production and welfare losses. This paper investigates the role of two market imperfections – lack of access to credit and lack of insurance – on these occupational choices among subsistence farmers in rural Kenya. Using a unique dataset of people’s subjective beliefs about the distribution of returns and entry costs for the set of income-generating activities available to them, I estimate a random utility model of occupational choice and find evidence that risk, and the lack of insurance, prevents entry into high profit but high variance occupations. Calculations of the compensating variation associated with changes in the variance, or riskiness, of these occupations suggest that income losses from this market failure can be a considerable fraction of expected earnings. In contrast, the relative size of entry costs appears to have no effect on people’s decisions, suggesting that access to credit does not constrain choice for the occupations available in the study area. JEL Codes: D52, D84, D92, O16 ∗I am grateful to Pascaline Dupas, Robert Jensen, Adriana Lleras-Muney, and Jonathan Robinson for helpful comments on this draft. I thank Sarah Green, Kathy Nolan, and Kim Siegal for research assistance in the field. All errors are my own. †Wesleyan University Department of Economics, email: [email protected]