Set Up a Pool

[yasir]

Dear All,
I am new to Eviews 7. I have reading the Eviews User Guide and this forum but still I am unable to set a pool for the the attached data. My data has countries, years, and different variables such as gdp, pci, distance etc. Now I was able to transform my data into Panel but I cannot un stack it to use in pool. I am attaching both the Eviews and Excel format files. I want to have separates series for every variable, e.g. trade_Arg, trade_JAP, gdp_Arg, etc. Can anyone help please?

Secondly, while doing panel estimation, when I try to use the cross section fixed effects, it gives me an error "Near Singular matrix". What does this mean?

Thanks in advance.


Yasir

[EViews Esther]

You can create a new workfile page by breaking series of interest (e.g. trade gdp) into multiple series based on the cross-sectional ids:

Code: Select all

pageunstack var01 id01 @ trade gdp

Which variables did you use for your regression? The error message (singularity) seems to imply that your regressors are perfectly collinear.
Also, please note that you are not going to be able to estimate the fixed effects model if at least a couple of your variables do not vary within cross-sections.

[yasir]

Dear Esther,
Many thanks for the prompt reply. I am writing after a long time because I went back to reading the books and stuff to clear the concepts of estimation. The problem was because Fixed Effects cannot be performed when we have data that does not change, in my case, the distance, language, religion etc. Once I removed these, I easily performed Estimation with Fixed Effects. But I still have a problem, it is mentioned in many research papers that perform a Hausman Test to find which estimation method suits your data, random or the fixed effects. I performed the test and I got these results but I have no clue what is going to determine whether to use Fixed Effects or the Random Effects Model. Can you help please?


Correlated Random Effects - Hausman Test
Equation: EQ01
Test cross-section random effects

Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.

Cross-section random 9.495735 4 0.0498


Cross-section random effects test comparisons:

Variable Fixed Random Var(Diff.) Prob.

GDP 0.563530 0.590019 0.000767 0.3387
EXCHANGE -0.158280 -0.087400 0.004725 0.3025
CPI -0.079056 -0.074239 0.001445 0.8992
TRADEGDP 0.040050 0.051622 0.000262 0.4749


Cross-section random effects test equation:
Dependent Variable: TRADE
Method: Panel Least Squares
Date: 05/20/12 Time: 04:33
Sample: 2000 2010
Periods included: 11
Cross-sections included: 107
Total panel (unbalanced) observations: 1135

Variable Coefficient Std. Error t-Statistic Prob.

C -10.02980 2.067527 -4.851108 0.0000
GDP 0.563530 0.046451 12.13180 0.0000
EXCHANGE -0.158280 0.084075 -1.882596 0.0600
CPI -0.079056 0.077077 -1.025684 0.3053
TRADEGDP 0.040050 0.069719 0.574450 0.5658

Effects Specification

Cross-section fixed (dummy variables)

R-squared 0.895219 Mean dependent var 17.61190
Adjusted R-squared 0.883963 S.D. dependent var 2.405221
S.E. of regression 0.819320 Akaike info criterion 2.531995
Sum squared resid 687.3965 Schwarz criterion 3.024345
Log likelihood -1325.907 Hannan-Quinn criter. 2.717969
F-statistic 79.53385 Durbin-Watson stat 1.129236
Prob(F-statistic) 0.000000


Secondly, just a question of Econometrics,

"There is, of course, a problem with FEM. We cannot directly estimate variables that do not change over time because inherent transformation wipes out such variables. Distance and dummy variables in our aforesaid models are such variables. However, this problem can easily be solved by estimating these variables in a second step, running another regression with the individual effects as the dependent variable and distance and dummies as independent variables"
Quote: Mohammad Mafizur Rahman, A Panel Data Analysis of Bangladesh’s Trade: The Gravity Model Approach.

What are individual effects and how to estimate them? Any ideas?

Thanks


Yasir