quarterly effects

[lemongrass]

Hi all,

I have to estimate a regression with 12 countries over 40 quarters. I've got a dated pool. If better, I could also make a dated panel data set.

I have an OLS estimation with Coef covariance method white period and cross-section fixed effects.
Now I want to implement quarterly effects, not period fixed effects.
I used @expand(@quarter, @drop(1)). The problem, Eviews take only quater 2, 3 and 4... note 39 quaters
-> how can I "tell" Eviews that it should estimate quarterly effects for 39 quarters?

THANKS for your help!!

[startz]

Do you mean there are only three dummy variables created or do you mean that the dummy variables only have values for the first three quarters?

[lemongrass]

Hi,

there are only three dummies created. I got @quarter-2, @quarter-3 and @quarter-4
I want dummies for quarter 2,3,...,40. For each quarter in all years, excluded the first quarter in my regression.

THANKS!

[startz]

Hi,

there are only three dummies created. I got @quarter-2, @quarter-3 and @quarter-4
I want dummies for quarter 2,3,...,47. For each quarter in all years, excluded the first quarter in my regression.

THANKS!

If you have quarterly data and you want a dummy for each quarter, then those are period fixed effects.

[lemongrass]

Hi,

thanks for your very fast reply :)

I've tried already period fixed effects. It works and I get no error message. But I wanted period fixed effects but not for quarter 1. I want to exlude the first quarter. Is this possible?

THANKS

[startz]

I suspect that EViews is already dropping one period, but I'm not sure about that.

In any event, you might try

Code: Select all

@expand(@date,@dropfirst)

[EViews Gareth]

The built in effects are deviations from mean. The regression is the same. If you really want the non-deviation form you can use Startz's suggestion.