Difficulty in getting results in a DID model

[Anushka Chauhan]

Hi, I want to use the DID model for 67 firms for 18 years. I am using the fixed effects model. I also wanted to use time fixed effects. And due to heterogeneity and cross-sectional dependence among panels, I also used xtscc command. But I am not getting any result. I need to know what I am doing wrong. Please let me know.

[George Ford]

What period does the treatment occur?

Do you have a dummy marking treated units?

Do you have a dummy marking the treatment period?

You could start with something like this if all treatment occur at the same time

reghdfe Y c.treated#c.post X, absorb(firmid time) cluster(firmid)


xtscc would work too if you need the DK standard errors. With it, you'd need a time FE specified and include fe as an option.