Objective: This notebook replicates the main findings of Abrevaya (2002) and provide code for quantile regressions in RStudio. In his paper, Jason Abrevaya studies the impact of demographics and maternal behavior on various quantiles of the birthweight distribution. The question is relevant because high costs and long-term effects -medical and economic- are associated with low-birthweight babies. Thus, the identification strategy allows to focus on the lower end of the birthweight distribution to quantify drivers of low-birthweight. Using data on births in the US in 1992 and 1996, he mainly highlights that many covariates have larger effects at the lower quantiles and lower effects at the higher quantiles. According to the author, the OLS can seriously under-estimate the effects at lower quantiles. Finally, results don’t seem to be driven by state effect.

Objective: This notebook replicates the main findings of Abrevaya (2002) and provide code for quantile regressions in RStudio. In his paper, Jason Abrevaya studies the impact of demographics and maternal behavior on various quantiles of the birthweight distribution. The question is relevant because high costs and long-term effects -medical and economic- are associated with low-birthweight babies. Thus, the identification strategy allows to focus on the lower end of the birthweight distribution to quantify drivers of low-birthweight. Using data on births in the US in 1992 and 1996, he mainly highlights that many covariates have larger effects at the lower quantiles and lower effects at the higher quantiles. According to the author, the OLS can seriously under-estimate the effects at lower quantiles. Finally, results don’t seem to be driven by state effect.