Dear Stata Community,
I'm running an IV regression with ivreghdfe, the unit of observation is a patient, the endogenous variable to be instrumented are dummies for identities of the doctor treating the patient, and the instruments are dummies for identities of the doctors who are on duty. The region is a just-identified setting. Since the data contains patients from different hospitals, I also control for hospital fixed effects. When I use ivreghdfe to run the IV regression, I find an unexpected pattern that some doctor dummies omitted as the reference group have standard errors, but some doctors dummies that are not omitted don't have a standard error (see details in the screenshot below). Might anyone knows what's causing this issue and how I may fix it? Thanks in advance!
My code is "ivreghdfe outcome (AsgnProv_* = AvailableProv_*), a(hospital)", where AsgnProv_* and AvailableProv_* are dummies for treating and on-duty doctors, respectively, and hospital is hospital ID.
I'm running an IV regression with ivreghdfe, the unit of observation is a patient, the endogenous variable to be instrumented are dummies for identities of the doctor treating the patient, and the instruments are dummies for identities of the doctors who are on duty. The region is a just-identified setting. Since the data contains patients from different hospitals, I also control for hospital fixed effects. When I use ivreghdfe to run the IV regression, I find an unexpected pattern that some doctor dummies omitted as the reference group have standard errors, but some doctors dummies that are not omitted don't have a standard error (see details in the screenshot below). Might anyone knows what's causing this issue and how I may fix it? Thanks in advance!
My code is "ivreghdfe outcome (AsgnProv_* = AvailableProv_*), a(hospital)", where AsgnProv_* and AvailableProv_* are dummies for treating and on-duty doctors, respectively, and hospital is hospital ID.
