I have a dataset on buyers and sellers and their transactions for multiple years. It looks like the following:
I want to estimate regressions of the form:
for each year, while storing the estimates (preferably in a common file).
As of now, I am running regressions of the following type:
Short of manually doing this year by year and storing estimates, I am thinking of doing something of the form:
and adding a command from the estimates store class of commands.
How can I save the estimates to the same external file, where the estimates from each regression corresponding to different years are matched by the variable names.
This way cannot work, because once I use parmby, the entire original dataset is deleted, and I will be unable to run multiple regressions. Any help is much appreciated.
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input float(Transaction Buyer Seller) str4 Year 121 1 2 "1979" 342 1 3 "1979" 342 1 4 "1979" 55 2 1 "1979" 443 2 3 "1979" 2324 2 4 "1979" 324 3 1 "1979" 55 3 2 "1979" 334 3 4 "1979" 55 4 1 "1979" 33 4 2 "1979" 22 4 3 "1979" . . . ".." . . . "." . . . "." . . . "." . . . "" . . . "" 213 1 2 "1995" 2141 1 3 "1995" . . . "." end
I want to estimate regressions of the form:
Code:
xi: regress Transactions i.Buyer i.Seller
for each year, while storing the estimates (preferably in a common file).
As of now, I am running regressions of the following type:
Code:
parmby " regress Transactions i.Buyer i.Seller ,label norestore
Code:
levelsof year, local(i)
foreach j of local i {
parmby " xi: regress Transactions i.Buyer i.Seller ,label norestore
}
How can I save the estimates to the same external file, where the estimates from each regression corresponding to different years are matched by the variable names.
This way cannot work, because once I use parmby, the entire original dataset is deleted, and I will be unable to run multiple regressions. Any help is much appreciated.