panel data problem

[windyboo]

Hello,

I've found these data (http://forums.eviews.com/viewtopic.php?t=74&f=7) to illustrate my questioning (vs my set of data which is bigger).

Year Country GDP Unemployment Inflation
1990 USA 45,18521039 71,83384877 71,15749544
1991 USA 23,04601164 78,76719258 56,60289052
1992 USA 29,69415401 46,21706091 70,55974091
1993 USA 53,55888923 78,60050714 77,54319289
1994 USA 19,47274418 31,85681226 76,86574562
1995 USA 91,50116172 98,91007925 63,91959817
1996 USA 24,84979164 40,67420505 58,79510315
1997 USA 53,52123958 45,11429455 18,41374564
1990 UK 85,98416975 2,065408081 79,77445146
1991 UK 87,49778804 96,39854289 39,46303189
1992 UK 68,00752296 53,67362608 10,63008431
1993 UK 89,58282722 36,18780323 24,06415912
1994 UK 36,88112043 14,28218606 95,57600277
1995 UK 13,5179681 39,154493 75,34766054
1996 UK 44,41815401 69,83427487 46,24098279
1997 UK 15,8282288 63,36841282 26,53225662
1990 CAN 26,34920496 8,712059759 90,91049229
1991 CAN 5,549156494 23,6257285 61,45173048
1992 CAN 95,45774226 11,49571302 43,66396842
1993 CAN 15,2350978 6,890314237 82,82921702
1994 CAN 77,43670278 93,07397669 21,7994829
1995 CAN 71,14403694 92,53688834 78,51250523
1996 CAN 29,89135716 89,42495363 42,60706258
1997 CAN 84,83811372 73,75955156 34,12932157
1990 AUS 13,18672879 58,12718154 32,66319198
1991 AUS 90,86758398 67,20240997 37,55421588
1992 AUS 57,86657672 53,90878347 90,88773844
1993 AUS 53,38208457 20,96291398 57,33999587
1994 AUS 92,73748539 71,91930498 86,78513806
1995 AUS 68,54576615 79,14986101 17,88874022
1996 AUS 19,40075366 40,99287255 30,16125981
1997 AUS 26,18899509 96,10888824 54,91768383

My question is pretty simple.

I want to do a regression but my independent variables will be on time t-1 and my dependent variable will be on time t. I don’t want a pooled regression, so I have to give instruction to the software. In my case, the equation is: GDPt = Unemployment t-1 + Inflation t-1

For example, if we look at the GDP of USA in 1997, I want that the software looks at the USA unemployment and inflation in 1996.

How can I do that (i.e. a regression which chooses the right data like I describe)?

Thank you in advance.

[EViews Gareth]

Just enter your equation as:

Code: Select all

gdp unemployment(-1) inflation(-1)

[EViews Glenn]

Make sure you are structured as a panel workfile for the lags to work properly.

I don't understand your comment that you **don't** want a pooled regression.

[windyboo]

I don't understand your comment that you **don't** want a pooled regression.

I've read that if you are doing a regression on panel data as you're doing a regression in time-series data (without spefications that you have many companies and not only one), it would be a ''pooled'' regression. Maybe I was unclear.

If I follow the directives of the second discussion (http://forums.eviews.com/viewtopic.php?t=74&f=7), it should be ok to create my panel data file from an Excel sheet.

Thank you for your advices (EViews Glenn and EViews Gareth).