Augmenting a Mincerian earnings function with regional data we estimate both private and external returns to education in Turkey using Instrumental Variables, Ordinary Least Squares, Quantile Regression and Instrumental Variables Quantile Regression methods. Our results indicate a median external return between 1.5% and 2.3% for 2006–2009. There is some evidence supporting the skill-biased technical change hypothesis. External returns are uniformly higher for women. We point out some policy implications.