This paper studies the causal effect of education on income uncertainty using a broad measure of income which encompasses unemployment risk. To accomplish this, the variance of residuals from a Mincer-type income regression is decomposed into unobserved heterogeneity (known to the individuals...