How to enter data into a panel workfile.

[Cciobu89]

Sorry for that, this is the one.

[EViews Gareth]

Japan has two unemployment series and no primary_balance series.

[ca1ek]

Hello is there any tutorials how to work with panel?

[EViews Gareth]

http://www.eviews.com/Training/Panel.html

[sam0000]

Hei Gareth.

I'm using the second kind of data, but when I try to unstack, I receive the ID error (ID series do not uniquely identify workfile observations). Could you help me, please? The file goes attached.

Thank you!

[EViews Gareth]

What is your cross-section identifier?

[sam0000]

What is your cross-section identifier?

Empresa (company, in English)

[EViews Gareth]

You have 33 rows for each year, for each empresa value. Thus it cannot be structured as a panel.

[sam0000]

But is there a limit of rows? What do you suggest?

[EViews Gareth]

Panel data is defined by having a single observation for each data/cross-section combination. You don't have a single observation. You have 33.

[sam0000]

Ok, but there is something strange here. The structure of my file is the same as yours (second type, first page). Take a look:

Empresa = State (you have 8 states/rows, I have 473 companies/rows)
Variável = Variable (you have 6 rows, I have 33 variables/rows)
Yr1940 = 2012
Yr1941 = 2011
(...) (at this point you have 31 years/columns and I have just 10)

Ok, there is a difference in the number of rows and columns, but the structure is quite the same.

I'm sorry if my doubts are simple, but this is something new for me and I'm trying to follow your example to deal with my data. And I have to write an article for this weekend...

Thanks!

[EViews Gareth]

Ah, the missing information was that Variavel was the series name. I should have seen that, my bad.
But the steps you need to take are identical to that given in the example. The only tricky bit is that your variable names are a bit long. Here's a program that will do it all for you:

Code: Select all

wfopen e:\downloads\teste.xls colhead=1 pagestruct(freq=u, start=1) pagestack num? smpl if variavel = "TamanhoFirma1" variavel = "TF1" smpl if variavel = "TamanhoFirma2" variavel = "TF2" smpl @all alpha names = @left(@makevalidname(variavel), 8) pageunstack names empresa var01 @ id01 d id01 rename id01* *

[sam0000]

Gareth, thanks for your help.

Now I have a different doubt. I'm running this regression with panel data with 10 years of data and more than 300 companies. The regressions are OK and the results are quite interesting, but the I have JB 4309.864(0.0000), which means that the residuals are not normally distributed. I've seen that this is not necessarily a problem and I still can use the coefficients and the p-value. But, do I have to do other tests because of this result?

Another question: I'm trying to test heterocedasticity and multicollinearity, but when I go (Eq03, added file) in View / Residual Diagnosis, there is just the Histogram - Normality Test option available. How can I deal with this question?

Thank you!!

P.S.: Eq03 is the model I'm using, because it seems to be the best. File added.

[MJJ]

Hi Gareth

Have any idea how to insert data from excel attached to Eviews panel workfile? As you see in excel file, I have four variables in four sheets, cross-sections are companies, named from comp1 to comp482 in top row of each sheet ... I know too many cross-sections, so if Eviews can't handle it, do you have any suggestion about any other statistical packages able to handle such a huge data?? Thanks in advance.

[EViews Gareth]

That data is nowhere near being huge :)

Somewhat simple EViews 8 program will get it in for you:

Code: Select all

%filename = "c:\temp\need to form as panel data.xls" 'file to be opened %sheetnames = @tablenames(%filename) 'get the names of the sheets %sheet = @word(%sheetnames, 1) wfopen(type=excel, page=all) %filename range={%sheet} 'open first sheet pagestack(page=allstk) comp? 'stack into a panel alpha compname = "comp" + @str(var01) 'create an alpha series containing company names pagestruct @date(date) compname 'structure the page by date and the company names rename comp {%sheet} 'clean up d var01 id01 date 'loop through remaining sheets, importing the variables for !i=2 to @wcount(%sheetnames) %sheet = @word(%sheetnames, !i) pageload(type=excel, page={%sheet}) %filename range={%sheet} pagestack(page={%sheet}stk) comp? alpha compname = "comp" + @str(var01) pagestruct @date(date) compname rename comp {%sheet} d var01 id01 date copy {%sheet}stk\{%sheet} allstk\{%sheet} next 'delete all the excess pages (uncomment below lines) 'for !i=2 to @wcount(%sheetnames) ' %sheet = @word(%sheetnames, !i) ' pagedelete {%sheet} ' pagedelete {%sheet}stk 'next pageselect allstk