EViews.com

Hi,

I am having some issues importing an excel data file into Eviews, It is panel data from 1997-2010 with 17 cross-sections, I just can't get eviews to recognize the dates correctly, or import all of the data.

Can anybody help me out with this?

Thanks,
Alix

I don't think you have 17 cross-sections. Either it is a three-way panel, or you need to create more cross-secions than you currently have. There is currently no way to uniquely identify each row based on a date and a cross-section.

I don't understand what I need to do to make it ready for importing into Eviews, how would I do either of those options?

It isn't an EViews issue per se. You need to think about how this data is a panel.

I understand, how would you advice editing the data so that it is compatible? How does the format need to be changed.

It appears that I have the year id (year). But I have 2 cross-section id (importcountry_ and exportcountry_). It appears that, somehow, I have an index, say 3-2005, and it points to two different observations. That’s why the panel set up fails.

How do you suppose I change this so that each id points to one observation?

Sorry if this is confusing, I am fairly new to Eviews, or really any statistical software.

Thanks for your help.

Well I can't really advise you on how you want to think about your data. It seems you have two options, you either have a 3-dimensional panel (one time series, two cross-sections), or you have a 2D panel (one time series, one cross-section).

EViews will handle 3D panels fine, in that you can import it in and tell EViews you have two cross-sections. However some of the estimation and statistical procedures rely on you having only two dimensions, not 3.

It seems you could create a 2D panel simply by adding a new column that identifies the cross-section component. Some sort of code (off the top of my head, you could use something like GER_FRA, GER_UK etc..).

Makes Sense! Thanks much for the help - I wasn't understanding the problem.