I study the effect of a conditional cash transfer program in Mexico on early marriage. The program provided monetary benefits to households, conditional on children’s school attendance. Leveraging on the staggered implementation of the program, I find that expos- ure to the conditional cash transfer increased girls’ probability of marriage. After five years of exposure to the program, beneficiary girls were, on average, 7 p.p more likely to be married than the control group. I find no effect for boys. These findings contrast with the previously documented positive effects of the program on education, which is usually associated with decreases in child marriage. I reconcile the simultaneous increase in marriage and education in a conceptual framework wherein agents treat marriage as a normal good. Finally, I test whether marriage responds positively to income by exploiting the variation in household composition and find that non-eligible children in beneficiary households - who were only exposed to the increase in household income - were between 10 and 18p.p more likely to be married than their counterparts in non-treated villages.