Dear Stata users,
I am working with panel data. I am running regressions for each id with:
bysort id: regress x y, robust
The regressions run fine and the results are displayed. I use Stata 17 SE and have 298 ids with 260 observations for each regression in this case. My ultimate interest is in the percentage of all regressions that yield a positive constant.
My first step for a solution was to store each regression output in one combined external file. My understanding from reading in the forum is that eststo and esttab are suitable commands to achieve the objective.
However, when running:
bysort id: eststo: regress x y, robust
I get the error message: "system limit exceeded
you need to drop one or more models". The regressions will then run but not store.
I am now looking for guidance how to store the 298 regressions or achieve the ultimate objective in a different way.
I have used a loop with outreg2 when the data is in timeseries format to extract the results, which works fine, but creates a lengthy cycle of computations (in a second step I am running these regression for multiple different subsets of the 260 observations). Is there an obvious reason what system limit is exceeded and how to remedy the situation? Should I instead try to store and insert the specific regression output (i.e. the constant) in a seperate variable? Any help is much appreciated.
best,
Daniel
I am working with panel data. I am running regressions for each id with:
bysort id: regress x y, robust
The regressions run fine and the results are displayed. I use Stata 17 SE and have 298 ids with 260 observations for each regression in this case. My ultimate interest is in the percentage of all regressions that yield a positive constant.
My first step for a solution was to store each regression output in one combined external file. My understanding from reading in the forum is that eststo and esttab are suitable commands to achieve the objective.
However, when running:
bysort id: eststo: regress x y, robust
I get the error message: "system limit exceeded
you need to drop one or more models". The regressions will then run but not store.
I am now looking for guidance how to store the 298 regressions or achieve the ultimate objective in a different way.
I have used a loop with outreg2 when the data is in timeseries format to extract the results, which works fine, but creates a lengthy cycle of computations (in a second step I am running these regression for multiple different subsets of the 260 observations). Is there an obvious reason what system limit is exceeded and how to remedy the situation? Should I instead try to store and insert the specific regression output (i.e. the constant) in a seperate variable? Any help is much appreciated.
best,
Daniel
Comment