topic Panel regression error in Statistical Procedures

Panel regression error

firatyilmaz — Thu, 20 Jun 2019 15:56:49 GMT

Hi everyone,

I have a panel data with missing values. When I try to estimate a fixed or random effects regression, the proc panel procedure returns with an error like this: "there are missing values in the TS variable". So I can't make an estimation. Can you help me how to solve this problem? I'm using SAS OnDemand for Academics by the way.

Re: Panel regression error

arthurcavila — Sat, 22 Jun 2019 04:37:39 GMT

Can you remove the missing observations? Maybe something like:

PROC PANEL data=your.data(where=(TS~=.));

*...(your model code);

RUN;

Re: Panel regression error

firatyilmaz — Sat, 22 Jun 2019 09:38:45 GMT

Hi,

Thank you for your answer. I added the code you suggested to the end of the proc panel statement, but this time it returned with two errors: "Variable TS is not on file WORK.GAS" and "No data set open to look up variables". My data actually an open access stata file(.dta) so I used the following code to import it to SAS:

%web_drop_table(work.gas);

filename gas "/home/u37951485/gasoline4.dta";
run;

proc import datafile=gas
     dbms=dta
     out=work.gas;
     run;
     
%web_open_table(work.gas);

My panel procedure is:

proc panel data=work.gas (here your suggested extension);
     id country year;
     model lgasconspc=lgasp co2int annual_temp /fixone;
     run;

The interesting thing is that I can estimate the model with proc reg without any errors. İn Stata, I also can estimate the model even though it is an unbalanced panel without any problems. I have read on SAS support that proc panel procedure would simply ignore the misssing values and estimate the model. But it does not. What is the problem or what am I doing wrong?