Does large-scale public primary school construction improve educational mobility in developing countries? No credible evidence on this. We estimate the effects of 61k primary schools in Indonesia. Short answer: improved mobility for disadvantaged sons, but not for daughters. Why? #Econsky