Rates of return to education in Sri Lanka

This paper explores the rates of return to education in Sri Lanka across the sexes and different types of employment during 2009/10. The endogeneity bias suggests that education may be associated with other characteristics such as ability and family background - excluding such attributes could lead to biased estimates of the return to education. To deal with this form of bias, the paper uses an Instrumental Variables (IV) approach to measure education. Our results reveal that the OLS estimates are downward biased, in comparison to the IV estimates. Further, we add to the literature by exploring the return to education across levels of education - primary, secondary and tertiary - and how it varies across sexes and types of employment. Finally for the self-employed, we analyse the association between inputs (land, raw materials, labour and capital), education and output using a household-level production function - for those in agriculture and non-agricultural activities. The results suggest that education at higher levels have a positive association with non-agricultural output, but not with agricultural output.