Insufficient observations

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

eviewsnc
Posts: 7
Joined: Tue Nov 10, 2015 4:21 am

Insufficient observations

Postby eviewsnc » Tue Jul 17, 2018 3:05 am

Dear all,

I am trying to estimate a fertility curve, using the data in the attached workfile.

I am trying to estimate the following equation using least squares (other estimation methods also do not yield results).

equation eq1.ls fertility=c(1)*(c(2)/c(3))*(c(3)/age)^(3/2)*exp(-(c(2))^2*(c(3)/age+age/c(3)-2))

I keep getting an error message saying I have insufficient observations, which I don't understand. Any assistance would be greatly appreciated.
Attachments
Fertility_eviews.WF1
(9.81 KiB) Downloaded 224 times

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Insufficient observations

Postby startz » Tue Jul 17, 2018 5:57 am

EViews starts estimation with whatever is stored in the c vector. In your case, this is causing some division by zero. Try saying

Code: Select all

c = 1
before you start the estimation.

eviewsnc
Posts: 7
Joined: Tue Nov 10, 2015 4:21 am

Re: Insufficient observations

Postby eviewsnc » Tue Jul 17, 2018 6:25 am

startz wrote:EViews starts estimation with whatever is stored in the c vector. In your case, this is causing some division by zero. Try saying

Code: Select all

c = 1
before you start the estimation.


Many thanks Startz, works perfectly.

eviewsaoa
Posts: 7
Joined: Tue Jul 17, 2018 7:59 am

Re: Insufficient observations

Postby eviewsaoa » Tue Jul 17, 2018 8:22 am

Dear all,

I've encountered the same issue, so your answer was very useful - many thanks for that.

Based on the workfile attached in this thread, I have a question concerning the starting point values. I have set the following equation and starting point values, which should be close enough to the actual curve:

param c(1) 1 c(2) -.15 c(3) 30 c(4) 8

equation eq1.ls fertility=2*c(4)^(-1)*exp(-(age-c(3))/c(4))/((1+exp(-(age-c(3))/c(4)))^2*(1+exp(-c(1)*((age-c(3))/c(4))-c(2)*((age-c(3))/c(4))^3)))


However, when I change these values, I do not get any changes in the estimated coefficients (which I would expected). Also, even if I close and open the program, I get convergence all the time in the same number of iterations. I wouldn't have expected this -maybe I'm thinking in terms of the "Solver" addin of Excel, and that's confusing me (relatedly, the estimation I get through Solver by minimizing the SSR is much closer to the actual figure compared to what I get in Eviews).

Any help would be highly appreciated.

Thanks in advance.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 13 guests