Hi all,
My key independent variable of interest can take one of four values (1,2,3,4) and there is serious bunching: 80% of all observations have value=3 with the rest of the observations being relatively distributed among the other 3 values.
The dependent variable can take the values of positive real numbers and does not suffer from bunching.
I estimate a two-way fixed effects regression of the form:
My question is whether there are any fancy estimation models I should be aware of which can better handle the fact that my data is bunched at one particular value for the independent variable?
Thanks in advance,
Rhys
My key independent variable of interest can take one of four values (1,2,3,4) and there is serious bunching: 80% of all observations have value=3 with the rest of the observations being relatively distributed among the other 3 values.
The dependent variable can take the values of positive real numbers and does not suffer from bunching.
I estimate a two-way fixed effects regression of the form:
HTML Code:
xtreg y i.x i.year, fe vce(robust)
Thanks in advance,
Rhys
