alpha series and cross-sectional data

[deroekonom]

hello,

I imported data from excel with three different observations for 25 regions.

However, I also have to import data where the 25 regions are cross-sectional linked. I.e there are two 25x25 matricices which I want to import into eviews too.

But I'am not sure how to do that. Maybe somebody can help me. I'am not sure but maybe the solution is to define cross-section
identifiers?

Thx for help.
doe

[deroekonom]

hello again,

I forgot to mention that one of the 25x25 matrices may be generated by one of the values assigned to the regions.

However, the second matrix has nothing to do with the values already assigned to the regions. The second matrix consists of values which are exogenously given.

Thx.

[EViews Gareth]

I've read our posts a couple of times, but I'm having a real problem understanding what you're trying to do

[deroekonom]

hello,

I'am sorry if I confused you.

My meaning was:
I have vales for every region in my sample, e.g.:

Florida: 25
New York: 50
California: 100

Now I want to have a matrix which shows the ratios of all the regions in comparion to all other regions; something like that:

Florida New York California

Florida 1 0.5 0.25

New York 2 1 0.5

California 4 2 1

I don't know how to do get that matrix in eviews.
Thx your help.

[EViews Gareth]

No easy way to do it, but with a little bit of work you can create it:

(note in this example I am first creating a dataset with your 3 datapoints. You could expand the matrix creation bit to work on your fuller dataset).

Code: Select all

'create data
create u 3
series values
smpl 1 1
values = 25
smpl 2 2
values = 50
smpl 3 3
values = 100

smpl @all


'calculate matrix
matrix out = @outer(values,values)
vector(3) ones = 1
matrix denom = @outer(values,ones)
denom = @emult(denom,denom)
matrix results = @ediv(out,denom)
d out denom ones