Cross-Section Effects

[Apple Tan]

Hi,

Using Eviews7.

I have got regressions results generated using Eviews. It has "Cross-Section Effects". However I seldom use Eviews, and would like to know what "Cross-Section Effects" is. Is it the co-efficients on the dummies? If yes, 1 of the dummies should have been dropped due to multicollinearity. But I note the model has interaction terms between the dummies and the regressor.

Refer to attachment for details.

How do I generate "Cross-Section Effects" in Eviews?
What about in Stata?

Thank you

[EViews Gareth]

They are cross-section deviations from the overall intercept mean. Which is one method of interpreting fixed effects.

[Apple Tan]

"They are cross-section deviations from the overall intercept mean. Which is one method of interpreting fixed effects."

Hi EViews Gareth,

Thank you for the reply. But i still don't quite understand what is "cross-section deviations from the overall intercept mean." is there any way I can read up on this?

Do you know how I can generate it in Eviews or Stata?


Thank you

[EViews Gareth]

If you think about having a dummy variable for each cross-section, take the mean value of the coefficients on the dummy variables, then take each coefficient's difference from that mean. That is what is shown.

What do you mean by "generate it in EViews or Stata"? Generate what?

[Apple Tan]

If you think about having a dummy variable for each cross-section, take the mean value of the coefficients on the dummy variables, then take each coefficient's difference from that mean. That is what is shown.

What do you mean by "generate it in EViews or Stata"? Generate what?

Thank you for the reply.

1) I mean how do I generate the cross section effects for the dummy variables in Eviews & Stata?

Thank you!

[Apple Tan]

If you think about having a dummy variable for each cross-section, take the mean value of the coefficients on the dummy variables, then take each coefficient's difference from that mean. That is what is shown.

What do you mean by "generate it in EViews or Stata"? Generate what?

Hi

"If you think about having a dummy variable for each cross-section, take the mean value of the coefficients on the dummy variables, then take each coefficient's difference from that mean. That is what is shown."

Sorry I really don't get it.
If we put in the dummy variables, 1 will be dropped due to multi-collinearity.

So how I find the mean= coeffficient of dummy variables/ number of dummies

and the cross-section effects of all 3 dummies?


Thank you

[EViews Gareth]

You can estimate an equation with no constant and a full set of dummies (i.e. you don't need to drop a dummy if you have dropped the constant). If you do that, you can then take the mean of the dummy coefficients, and call that the constant. You can then show the difference between each dummy coefficient and their mean.

[Apple Tan]

You can estimate an equation with no constant and a full set of dummies (i.e. you don't need to drop a dummy if you have dropped the constant). If you do that, you can then take the mean of the dummy coefficients, and call that the constant. You can then show the difference between each dummy coefficient and their mean.

then how do I generate the cross section effects for the dummy variables in Eviews & Stata?

Thanks

[EViews Gareth]

generate the cross section effects for the dummy variables in Eviews & Stata?

You're going to have to explain what you mean.

[Apple Tan]

generate the cross section effects for the dummy variables in Eviews & Stata?

You're going to have to explain what you mean.

Hi

Thank you for the reply

In the attachment of my 1st post, there are the values of the cross-section effect for each of the dummy variables. Since this table was given to me, I have no idea how to obtain the "values of the cross-section effect for each of the dummy variables" in Stata.
I have figured it out in Eviews: View → Fixed/Random effects - > Cross-section effects.