BookmarkSubscribeRSS Feed
Tasneem
Calcite | Level 5

Basically, I want to run a panel regression on firm-years using industry and year fixed effects, and also get robust standard errors. I get the following in my SAS log when I use proc panel (sample code provided below):

 

**********************************************************************************************************************;

NOTE: The transformed regression does not have full rank. Be aware of possible multicollinearity
and/or identification problems before using the FixOneTm method results.
WARNING: Unable to scale time series group 2 for variable CSIndex. Values are all missing, or series
is nearly constant.
WARNING: Unable to scale time series group 8 for variable CSIndex. Values are all missing, or series
is nearly constant.
WARNING: Unable to scale time series group 12 for variable CSIndex. Values are all missing, or
series is nearly constant.

*********************************************************************************************************************;

 

The warning keeps getting generated with increasing "time series group" numbers until the results finally come out. My dataset is unbalanced (some firms have more years than others) but I do not think that is causing the problem. What do the note and warnings mean?

 

Here is some sample code I run:

 

/* "gvkey" is cross-sectional id & "fyear" is time-series id */

proc sort data=a; by gvkey fyear; run;

proc panel data=a;
    id gvkey fyear;
    class x1 x2;
model y = x1 x2 x3 ind1-ind40 / fixonetime hccme=1; run; /* x1 and x2 are categorical variables so I put them in a class statement */
/* ind1-ind40 are industry dummies for my data that I generated already*/

Thanks for you help! [SAS 9.4 on Windows 10]

 

1 REPLY 1
pink_poodle
Barite | Level 11

There are probably variables in the model with data that is similar to other variables, producing multicollinearity. CSIndex values are monotonous or missing.

 

ranking - sorting numerical or ordinal variables in ascending order 

 

multicollinearity - when one predictor variable in multiple regression model can be linearly predicted from the others

 

problem of multicollinearity - the coefficient estimates of the multiple regression may change erratically in response to small changes in the model or the data

 

some causes of multicollienarity:   inclusion of a variable which is computed from other variables in the data set; the repetition of the same kind of variable. 

 

scaling - normalizing data to a common range

 

See also,

 

The Panel procedure: Unbalanced data

Ready to join fellow brilliant minds for the SAS Hackathon?

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. Registration is now open through August 30th. Visit the SAS Hackathon homepage.

Register today!
What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 1 reply
  • 1749 views
  • 0 likes
  • 2 in conversation