Practice ahead of theory: instrumental variables, natural experiments and inductivism in econometrics

This paper argues that the field of econometrics divides into camps of sinners and preachers. Specifically, I look at two sins condemned by the preachers: the making of causal background assumptions when identifying parameters for estimation and the estimation of parameters that have no or little theoretical significance. I argue that practitioners in the natural experiments literature happily commit these sins and that this is for the good of the discipline.