Difference-in-differences is roughly 25% of all papers published in 2018 at the NBER working paper series and around 17% of all papers published in top 5 economics journals, but do you know what is even more popular? Instrumental variables. It constitutes over 30% of either!
I guess it’s not really a surprise to see it so dominant given it’s able to extract some version of an average causal effect from observational data. It can be used to solve all sorts of thorny policy-relevant problems, from figuring out the price elasticity of demand (a crucial parameter for maximizing profits, setting taxes or even deciding whether to regulate something at all), to measuring the repercussions of juvenile incarceration on high school completion and adult crime, to learning whether low quality hosts on Airbnb are imposing negative externalities on others by driving away customers. Just as Archimedes said “Give me a lever long enough and a fulcrum on which to place it, and I shall move the worl…
Watch with a 7-day free trial
Subscribe to Scott's Substack to watch this video and get 7 days of free access to the full post archives.
