Re: [Gretl-users] a problem about panel data
On Sun, 1 Dec 2013, Allin Cottrell wrote:
On Sun, 1 Dec 2013, Qi Shi wrote:
> when I replicate Example 11.3 of Wooldridge's advance book(the second
> edition), i encounter a problem. To run the pooled OLS, I write
>
> open airfare.dta
> genr lfare_1=lfare(-1)
> diff lare lfare_1 concen
> ols d_lfare d_lfare_1 d_concen
>
> then, I get the outputs, which is different from the results of the Book.
Please post the results shown by Wooldridge, in as much detail as possible. I
don't have his book to hand at present.
Ah, maybe you don't have to do that. You will get wrong results from the
above becaause you did not define the dataset as a panel before doing
first-differencing. You need
open airfare.dta
setobs id year --panel-vars
...
Then you get:
Model 1: Pooled OLS, using 2298 observations
Included 1149 cross-sectional units
Time-series length = 2
Dependent variable: d_lfare
coefficient std. error t-ratio p-value
---------------------------------------------------------
const 0.0391942 0.00252693 15.51 1.12e-51 ***
d_lfare_1 -0.134426 0.0189432 -7.096 1.70e-12 ***
d_concen 0.105206 0.0344988 3.050 0.0023 ***
Mean dependent var 0.035789 S.D. dependent var 0.120723
Sum squared resid 32.57246 S.E. of regression 0.119134
R-squared 0.027014 Adjusted R-squared 0.026166
F(2, 2295) 31.85930 P-value(F) 2.25e-14
Log-likelihood 1629.799 Akaike criterion -3253.599
Schwarz criterion -3236.379 Hannan-Quinn -3247.320
Does this agree with Wooldridge?
Allin Cottrell