Page 1 of 1

Std errors for relationships btwn estimated coefficients

Posted: Mon Oct 19, 2015 11:37 pm
by gibson mudiriza
Hello,
I have estimated the following nls regression model
Coefficient Std. Error t-Statistic Prob.
C(1) 2.726114 0.131661 20.70563 0.0000
C(2) 0.103135 0.031590 3.264839 0.0012
C(3) 10.00232 3.221113 3.105236 0.0021
C(4) -4.055268 1.115459 -3.635514 0.0003
R-squared 0.480138 Mean dependent var 8.381672
Adjusted R-squared 0.475682 S.D. dependent var 0.393807
S.E. of regression 0.285155 Akaike info criterion 0.339665
Sum squared resid 28.45961 Schwarz criterion 0.383386
Log likelihood -56.12066 Hannan-Quinn criter. 0.357060
F-statistic 107.7517 Durbin-Watson stat 1.673422
Prob(F-statistic) 0.000000

From this model l could like to derive some relationships based on the estimated coefficients say: c(2)(1-c(3)). My question is: Is there a way to get standard errors for the new derived relationships to get the significance level? If so can someone give me advise on who to get them.

Thank you in advance

Re: Std errors for relationships btwn estimated coefficients

Posted: Tue Oct 20, 2015 4:29 am
by EViews Gareth
View->Coefficient Diagnostics->Wald Test.

The Wald Test will produce the standard errors for any relationship you type in.

Re: Std errors for relationships btwn estimated coefficients

Posted: Tue Oct 20, 2015 11:26 pm
by gibson mudiriza
Thank you Gareth.