Hello everyone,
Hope you're well.
First of all, I'm new to all of this. I'm struggling to identify where I'm going wrong in either setting up my xtset or in my xtreg [variables], fe.
I'm looking at the impact of constructing a sports stadium on property prices, constructing a hedonic pricing model (including things like beds, baths, garage...). I've run some difference-in-difference tests on the data, however I'm trying now to run some fixed effect regressions.
I've set it up (xtset) so my panel variable is the address (when encoded) and the time variable is the sales date (through a date() variable) - though this could be wrong?
The type of variables I have are byte for bed, bath, garage, as well as for my dummies (I don't know if that is a red flag or not?). Float for the date() sale date. Long for my encoded variables Address, type (Detached etc) and age (New build or not).
Subsequently, I've run my regression such that:
. xtreg LNPrice SQFT Trend i.Type i.Age Bed Bath Garage, fe robust
There are other variables to input, socioeconomic and distance etc, but I've also tried running the fixed effects regression with those included and it is giving me the same issue: all my variables end up omitted due to collinearity. The only variables which generate coefficients are SQFT and Trend.
If anyone has any idea on where I may be making a mistake, I'd be most grateful.
I feel my only other option would be to try the long method of setting up fixed effects through dummy variables, meaning I can run xtreg without the ', fe' at the end.
All the best.
Hope you're well.
First of all, I'm new to all of this. I'm struggling to identify where I'm going wrong in either setting up my xtset or in my xtreg [variables], fe.
I'm looking at the impact of constructing a sports stadium on property prices, constructing a hedonic pricing model (including things like beds, baths, garage...). I've run some difference-in-difference tests on the data, however I'm trying now to run some fixed effect regressions.
I've set it up (xtset) so my panel variable is the address (when encoded) and the time variable is the sales date (through a date() variable) - though this could be wrong?
The type of variables I have are byte for bed, bath, garage, as well as for my dummies (I don't know if that is a red flag or not?). Float for the date() sale date. Long for my encoded variables Address, type (Detached etc) and age (New build or not).
Subsequently, I've run my regression such that:
. xtreg LNPrice SQFT Trend i.Type i.Age Bed Bath Garage, fe robust
There are other variables to input, socioeconomic and distance etc, but I've also tried running the fixed effects regression with those included and it is giving me the same issue: all my variables end up omitted due to collinearity. The only variables which generate coefficients are SQFT and Trend.
If anyone has any idea on where I may be making a mistake, I'd be most grateful.
I feel my only other option would be to try the long method of setting up fixed effects through dummy variables, meaning I can run xtreg without the ', fe' at the end.
All the best.
Comment