Lead exposure and cognitive skills in a developing country : evidence from toxic sites in Indonesia

We provide causal evidence from a developing country context that children exposed to lead at a young age exhibit worse cognitive outcomes. We exploit variation in exposure to leadcontaminated toxic sites in Indonesia and estimate a two-way fixed effects model using variation in age at first exposure and variation in distance to the site. We find that children who were exposed in utero or up to 6 years old scored -0.48 s.d. lower in a test of numeracy and -0.36 s.d. lower in a test of general cognitive ability if they lived within 3 kilometers of the toxic site compared with those who lived beyond 6 kilometers. Confirming that the impact of lead exposure depends on distance, we find that the impact is halved for children who lived 3-6 kilometers from the site and were exposed before 7 years old. The learning penalty in numeracy is equivalent to about 3 years of primary schooling in Indonesia for those who lived within 3 kilometers and 1.4 years for those who lived 3-6 kilometers from the toxic site. The effect size indicates the need to prevent lead exposure, especially at younger ages.