This paper estimates the impact of the boom in international gold prices on child labor and schooling in Colombia. I first set up a simple agricultural household model of child labor and commodity prices hocks which guides the empirical analysis. Then, I use individual level information from the censuses of 1985, 1993 (when prices where stable) and 2005 (when prices surged) merged with regional data on gold production capabilities. I define Gold Boom as an interaction between regional gold production capabilities and the international price of gold. I find that child labor is increasing (0.3 standard deviations) and school attendance is decreasing (0.9 standard deviations) in the measure of gold boom. Accordingly, the gold boom decreases school attainment (0.2 standard deviations). This is consistent with the model when initial child labor is low and substitution effects dominate income effects. Finally, I find that the years of education of the head of the head of the household but not her ownership of assets mitigate the collateral effects of the gold boom.