I follow the fresh new strategy out of Acemoglu and Angrist (2001) and you may interpolate by the extending earlier data

I follow the fresh new strategy out of Acemoglu and Angrist (2001) and you may interpolate by the extending earlier data fifteen Inability to adjust the high quality problems in the first phase additionally the involved F analytics renders a giant difference between the current studies. Instead clustering, the newest F statistics to rise to