Page 1 of 1

Error in estimating an Exponential Trend

Posted: Fri Feb 21, 2014 6:10 pm
by vksf
I am using Eviews 8 and I am getting an overflow/error message when I am estimating an exponential trend. I have typed the following equation under the quick tab using LS. (r is the independent variable and t=1,2,... is the time trend) Please find my workfile attached.
r c(1)*exp(c(2)*t)

It would be great if someone could help with this problem

Re: Error in estimating an Exponential Trend

Posted: Fri Feb 21, 2014 6:22 pm
by EViews Gareth
You're missing an equals sign

Re: Error in estimating an Exponential Trend

Posted: Fri Feb 21, 2014 6:34 pm
by vksf
Thanks Gareth. Even with the = it does not work.
I've even tried typing ls r = c(1)*exp(c(2)*t) in the command window

Re: Error in estimating an Exponential Trend

Posted: Fri Feb 21, 2014 6:54 pm
by vksf
I redid the exercise and it says "Near Singular Matrix Error. Regressors may be perfectly collinear"

Re: Error in estimating an Exponential Trend

Posted: Fri Feb 21, 2014 9:46 pm
by startz
Try typing

Code: Select all

c=0.5
before you do the estimate

Re: Error in estimating an Exponential Trend

Posted: Tue Feb 25, 2014 4:45 pm
by vksf
Thanks but does this solve the singularity issue? What is the basis for selecting c?

Re: Error in estimating an Exponential Trend

Posted: Tue Feb 25, 2014 4:47 pm
by EViews Gareth
Just choose something other than zero. It should remove the singularity error.