This paper looks at the joint impact of labour market risk and selection into employment on returns to education estimates. The risk-adjusted returns to both high school and college for males are larger than unadjusted returns. For females, risk leads to an increase in returns to high school but to a "decrease" in the returns to college while correcting for selection into employment has large effects for females. The results suggest that failure to account for risk and selection into employment when calculating returns to education leads to biased estimates.