An article on the BBC website called “School odds stacked against summer babies, says IFS” caught my eye. Joshua D. Angrist and Alan B. Krueger published a famous paper in 1991, which used the age of school entry as an instrumental variable to measure returns to schooling. A few years later John Bound, David A. Jaeger and Regina M. Baker wrote an even more famous paper in the Journal of the American Statistical Association, exposing problems of weak instruments used in Angrist and Krueger’s work. The BBC article says that “August children are 20 per cent less likely to attend a top university”, which seems like a pretty striking result. Did the IFS repeat a 20-year old mistake or does the old instrument work much better in the UK?