Thread by @jmwooldridge

If in a staggered DiD setting I write an equation with a full set of treatment indicators by treated cohort and calendar time, and include c(i) + f(t) (unit and time "fixed effects"), would you still call that a "fixed effects" model?

If you answer "yes" then you should stop saying things like "there's a problem with the TWFE 'model'." The modeling is our choice; we choose what to put in x(i,t) when we write

y(i,t) = x(i,t)*b + c(i) + f(t) + u(i,t)

The phrase "TWFE model" refers to c(i) + f(t), right?

If x(i,t) = w(i,t) -- a single treatment indicator -- then the model might be too restrictive. But as I've shown in my DiD work, it's easy to put more in x(i,t) and estimate a full set of heterogeneous TEs. But I can (and should) still use the TWFE estimator.

Which gets back to my point that I'm not a fan of the phrase "TWFE model," as if there is some immutable model that we have no control over. The problem is our modeling, not the TWFE estimator. Do I blame the maximum likelihood estimator when I apply it to a poor model?

So if you see that someone has estimated the constant effect model by TWFE, emphasize they should also try making the model more flexible and then estimate the extended model by TWFE. This turns out to be the same as a pooled OLS, RE, and imputation method.

Just want to make sure we're all on the same page with the recent DiD literature. It should be uncontroversial to distinguish between the model and the estimation method. We can get pretty far with traditional estimation methods.

#metricstotheface