Panel: matching different series

For questions regarding the import, export and manipulation of data in EViews, including graphing and basic statistics.

Moderators: EViews Gareth, EViews Jason, EViews Steve, EViews Moderator

Wecon
Posts: 11
Joined: Sun Feb 28, 2016 10:59 am

Panel: matching different series

Postby Wecon » Tue Apr 05, 2016 1:06 pm

Dear all,

I have a panel dataset in EViews 8 which consists of several series: A, B, and C. I now wish to import another series X, which contains extra information on the observations already the first dataset. The problem is that series X does not match with the other series, that is, the workfile with series X and the workfile with series A, B, and C are both unbalanced panels. Nevertheless, I can make sure that X contains at least all the observations in A, B, and C. Is there an easy way to add X to A, B, and C?

Thanks!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 12573
Joined: Tue Sep 16, 2008 5:38 pm

Re: Panel: matching different series

Postby EViews Gareth » Tue Apr 05, 2016 1:16 pm

Too abstract to understand what you want to do.
Follow us on Twitter @IHSEViews

startz
Non-normality and collinearity are NOT problems!
Posts: 3577
Joined: Wed Sep 17, 2008 2:25 pm

Re: Panel: matching different series

Postby startz » Tue Apr 05, 2016 3:23 pm

Remember that EViews needs a way to know how to match observations. You may want to look at Match merge in the Help system.

Wecon
Posts: 11
Joined: Sun Feb 28, 2016 10:59 am

Re: Panel: matching different series

Postby Wecon » Wed Apr 06, 2016 12:48 am

@EViewsGareth

Sorry if the question wasn't clear, let me specify: I have an eviews workfile containing IMF-data on international investment positions (IIPs) (different series, such as assets, liabilities, portfolio investment, FDI, ...) for a panel of Eurozone member states. The time-dimension differs for each country, for example, Slovenian data starts in 2008Q1 and Spanish data in 1993Q1. All this data is expressed in euro's, but I want it to be in % of gdp (the IMF doesn't provide this ready-made apparently). So, my plan is to download a series with gdp-data from the IMF or Eurostat. However, before downloading the gdp-data, I would like to know if:
(1) I can just download data on gdp for the countries in my sample without having to specify for each country individually the exact time period that matches the time period for that country in my IIP file. So, can I just donwload gdp-data from 1993Q1-2015Q3 for all countries, import it to EViews, and let EViews match each gdp-(country-time)-observation with the corresponding (country-time)-observations on IIPs. How should I do this?
(2) If the previous does'nt work, would it work when I did download only the gdp-data that matches exactly, for each country, the time period in my IIP-dataset?

I'll certainly look into startz suggestion.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 12573
Joined: Tue Sep 16, 2008 5:38 pm

Re: Panel: matching different series

Postby EViews Gareth » Wed Apr 06, 2016 7:18 am

Still to abstract. But EViews will probably do what you want pretty easily.
Follow us on Twitter @IHSEViews

Wecon
Posts: 11
Joined: Sun Feb 28, 2016 10:59 am

Re: Panel: matching different series

Postby Wecon » Thu Apr 07, 2016 12:22 am

@EViewGareth

Another shot for clarification: Suppose I have two workfiles (IIP-workfile and gdp-workfile) with different samples (country-dimension is the same, but for the time-dimension the gdp-workfile may contain longer timeperiods for some countries). Now, I want to add the gdp-series from the gdp-workfile (with longer time periods for some countries) to the IIP-workfile and do some calculations (for example, one series in the IIP-workfile is called 'assets': I want to calculate "assets_pct=assets/gdp"). But for that calculation to be correct, I need the assets of Spain-1995Q3 to be divided by gdp of Spain-1995Q3 (and not some other (country-time)-observation), and the assets of France-2009Q2 have to be divided by the gdp of France-2009Q2. I just don't know how to make sure that EViews understands that it has to match the correct (country-time)-observation in the gdp series with the corresponding (country-time)-observation in the asset series.

I believe I could use the one-to-one match-merge, as suggested by startz with two identifiers: a country-identifier and a time-identifier. It probably is indeed quite simple, buth I'm sure that the first time using panel data in EViews is always a bit of a challange :wink: .

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 12573
Joined: Tue Sep 16, 2008 5:38 pm

Re: Panel: matching different series

Postby EViews Gareth » Thu Apr 07, 2016 7:36 am

It should just be a case of dragging and dropping. As long as both are panels with the same cross-section identifiers, EViews will figure out the rest for you.
Follow us on Twitter @IHSEViews

Wecon
Posts: 11
Joined: Sun Feb 28, 2016 10:59 am

Re: Panel: matching different series

Postby Wecon » Fri Apr 08, 2016 1:12 am

It did indeed. No need to use match merge. Thanks!


Return to “Data Manipulation”

Who is online

Users browsing this forum: No registered users and 20 guests