Replication codes of some papers.
It is common in the education literature to use state-level compulsory schooling requirements in the US as an instrument to estimate the causal impact of education on many different factors. The identification, using this method, relies on the different moments that these laws were implemented across states, which means that different birth cohorts in a state faced different schooling requirements. Implied by this is that we’re assuming that all other changes that occur across states during this period are uncorrelated with the law changes, educational improvements and the outcomes studied.
More recent research shows that this assumption is unlikely to hold, as examples we have; Card and Krueger (1992) showing that school quality improved more rapidly in the south, which had important effects on wages. Other examples also show that many different factors during this period improved schooling outcomes of black and southern students. Indicating that the estimates of increased schooling may be driven by different factor that had different effects across regions of the US. With this in mind it is important to allow for the birth effects to vary across the different regions of birth in order to understand the real returns to schooling. Allowing for this variation the researchers show that the results from commonly-used baseline specification are driven by differences between regions as opposed to variation within states over time.