Page 1 of 1

Outcomes from Eviews commands in a program

Posted: Thu Apr 29, 2021 12:24 pm
by remuct
Hello,

I'm performing a system estimation of several equations using the SUR method with common structural breaks. In order to combine a system model with structural breaks I wrote a program file following methods by Baltagi et al (2015) and Bai and Perron (1998).

To get the standard deviation of the residuals of each equation k, I used the scalar value from System Data Members : @se(k).

However this value is not the same as the S.E. of regression provided at the bottom of Eviews output for the k'th equation after the estimation.

Furthermore, when I ask the covariance matrix of the residuals using @residcov(k,k), the square root of the k'th diagonal element is yet a different value than the two preceding ones.

Shouldn't these three functions give the same number ? If not what are their differences?

Thank you for your response
Best

Re: Outcomes from Eviews commands in a program

Posted: Thu Apr 29, 2021 1:24 pm
by EViews Gareth
The reported kth equation SE should match the system.@se(k) value (and does in a quick test).

The residual covariance matrix will differ because of a degree of freedom correction.

Re: Outcomes from Eviews commands in a program

Posted: Thu Apr 29, 2021 2:45 pm
by remuct
Thank you Gareth, this was useful.