Count within Cross-sectional identifiers

[tchaithonov]

Hi Everybody,

I am trying to do a count of the number of cross sections within my pooled OLS model, and I have "market" as my cross-section ID. How should I do it? Please share your thoughts. Thanks.

Tchaithonov

[EViews Gareth]

Assuming you're working with a Pool object, you can do:

scalar count = poolname.@ncrossest

where poolname is the name of your pool object.

[tchaithonov]

I am actually not working on a pool object. I have a stacked dataset with a cross section identifier (mkt) and a time series identifier (week), and now I have set it up as a balanced panel. As I was told, I should be running panel (so, using LS instead of POOL) for the large numbers of cross sections I have. So, my series are not "pooled" together. Is there still a way to do so? Please let me know. Thanks.

[EViews Gareth]

Yep, you can use:

equation_name.@ncross

[tchaithonov]

I tried it, but eviews keeps complaining that it has a syntax error for the @ncross line. I could find @ncoef and @regobs in the manual/help file, but not @ncross.

[EViews Gareth]

What's the version number and build date of your copy of EViews (both available from Help->About EViews)

[tchaithonov]

I tried this on an Eviews 6 Apr 20 2007 build and an Eviews 5 Jan 20 2006 build.

[EViews Gareth]

Update your EViews 6.

[tchaithonov]

Neat! It works now. Thanks for your help.