Re-assessing the impact of Brexit on British fertility using difference-in-difference estimation

Difference-in-difference (hereafter DID) estimation is a tool for extracting quasi-causal estimates from observational data. As an analytic strategy, it has a number of advantages over traditional regression or covariate adjustment approaches. In a recent contribution to this journal, XXXXXXXX uses DID to examine the impact of Brexit on fertility dynamics among European countries and concludes that there are large and robust Brexit effects. The purpose of this comment is to elaborate on the complexities of DID estimation and the critical role of “research degrees of freedom” – the decisions that DID analysts must make to satisfy the underlying assumptions of the approach. Empirically, we use data from the World Bank Database to re-assess the "Brexit hypothesis" given a variety of assumptions and derivative samples. In the end, there is little support for the "Brexit hypothesis." It is hoped that this comment provides a template for better use of the tool in demographic research.