unbalanced/balanced data problem

franta311 · Postby **franta311** » Tue Nov 04, 2014 2:55 pm

Hello!

I am doing Fixed Effects and Pooled OLS regressions using panel data in EViews 7. In the regression output (7th row) I have: Total panel(unbalanced) observations: 414.
I would like to ask you: why is there "unbalanced", when if I structured my data as Dated panel? Also, why is there only 414 observations, should not there be 42 countries, 13 time period, 12 explanatory variables => (42*13*12) – 151(missing data) = 6401 observations?

Many thanks!
F.

EViews Gareth · Postby **EViews Gareth** » Tue Nov 04, 2014 3:43 pm

Some (many, by the sounds of it) observations have missing data.

franta311 · Postby **franta311** » Tue Nov 04, 2014 4:09 pm

yeah, 151 observations have missing data, but why it is written 414 observations, if I should have 6401? (42countries*13years*12variables).

Any idea, why it says total panel(unbalanced) in the regression output, even if my data is structured as Panel Data and thefore should be balanced?

EViews Gareth · Postby **EViews Gareth** » Tue Nov 04, 2014 4:27 pm

More than 151 observations have missing data.

franta311 · Postby **franta311** » Tue Nov 04, 2014 5:21 pm

how can you know? exacly 151 observations have missing data.

and what I would like to also ask you why does it say total panel(unbalanced) in the regression output, even if my data is structured as Panel Data and thefore should be balanced?? - (it is the 7th row in the regression output)

EViews Gareth · Postby **EViews Gareth** » Tue Nov 04, 2014 7:46 pm

Well without providing the data, I can only go by the scant clues you've given. But the fact that EViews dropped so many observations indicates that many observations had missing data.

It is unbalanced because cross-sections have missing values at different observations.

franta311 · Postby **franta311** » Wed Nov 05, 2014 5:19 am

I am sending my data as attachement

EViews Gareth · Postby **EViews Gareth** » Wed Nov 05, 2014 8:52 am

Provide the EViews workfile, with the equation.

franta311 · Postby **franta311** » Wed Nov 05, 2014 12:11 pm

sending as an attachment

EViews Gareth · Postby **EViews Gareth** » Wed Nov 05, 2014 12:52 pm

The workfile has 546 observations total. The equation uses 414 of them. So 132 are dropped. They are dropped unevenly from different cross-sections, hence the unbalanced nature of the estimation.

Note that observations are counted as rows of data, not individual data values.

franta311 · Postby **franta311** » Wed Nov 05, 2014 1:29 pm

ok. thanks.

Does the unbalanced nature effect in any way the results?

EViews Gareth · Postby **EViews Gareth** » Wed Nov 05, 2014 1:52 pm

franta311 · Postby **franta311** » Thu Nov 06, 2014 6:39 pm

ok, thanks for your help!

EViews.com