Falsifications are just as important as event studies
Today’s difference-in-differences is not our grandfather’s difference-in-differences. We have survived a war but we lived to tell. We learned that our greatest sin had the unthinkable comparison of treated groups to already-treated groups. We repented, and swore it would never happen again.
But as the dust cleared, and thousands of citations accumulated to a handful of econometricians, small little details that maybe weren’t as historically obvious are making their way to the top, and one lesser known little assumption I’d like to just bring your attention is called the no anticipation assumption, as well as point us back to basics, which is that falsifications can also be important tools in your arsenal, not just event studies.
Common assumptions
The most common assumption invoked in a difference-in-differences design is called the parallel trends ass…
ncG1vNJzZmibkarAorjIp51nq6WXwLWtwqRlnKedZL1wssClqqKemZiutbXOp6pmmaKaeqvB0q1kmqtdnrqxu9GtmKes