[Eviews] - VIF help

[coolsaad]

in eviews there are ''centered vif'' and ''uncentered vif''......which one to follow and why??? plz sum1 help

[vinux]

Which equation(model) you are referring? I haven't used Eviews for OLS. I guess the difference would be probably with intercept and without intercept. You could go with with the intercept ( probably the centred vif).

[coolsaad]

thnks....bt im still nt satisfied

[coolsaad]

Breusch-Godfrey Serial Correlation LM Test:


F-statistic 7.073483 Prob. F(1,26) 0.0132
Obs*R-squared 6.630024 Prob. Chi-Square(1) 0.0100



Test Equation:
Dependent Variable: RESID
Method: Least Squares
Date: 06/29/12 Time: 23:11
Sample: 1970 2000
Included observations: 31
Presample missing value lagged residuals set to zero.


Variable Coefficient Std. Error t-Statistic Prob.


EXPORT 0.234025 0.420332 0.556761 0.5825
CURRENTACCOUNT -0.298364 0.912781 -0.326873 0.7464
IMPORT -0.286585 0.505336 -0.567118 0.5755
C 33.33938 996.0048 0.033473 0.9736
RESID(-1) 0.650298 0.244509 2.659602 0.0132


R-squared 0.213872 Mean dependent var 2.35E-12
Adjusted R-squared 0.092929 S.D. dependent var 2408.256
S.E. of regression 2293.630 Akaike info criterion 18.46035
Sum squared resid 1.37E+08 Schwarz criterion 18.69164
Log likelihood -281.1354 Hannan-Quinn criter. 18.53574
F-statistic 1.768371 Durbin-Watson stat 1.295320
Prob(F-statistic) 0.165556

[vinux]

I used Eviews6. Mainly for GARCH and VAR models. But I don't remember VIF option.

Regarding the above test. F-test suggest that it is fail to reject the null hypothesis ( Prob(F-statistic) 0.165556). i.e. you can conclude that there is no serial correlation under this test.
(I would suggest you take look at correlogram also to confirm this,)