Page 1 of 1

Nonlinear S curve estimation with panel data

Posted: Wed Dec 15, 2021 7:22 am
by AH2021
Hi all,

I'm new to EViews, having used Stata for 10 years, so apologies if this is a beginners question.
I am struggling with estimating the following (S curve) equation:
Y = 1/ (C1+C2 * C3 ^X)
I have a panel of 57 countries with data over 1996 to 2019 and I want to pool the data to estimate the above equation.

I read the data into EViews as dated panel, with country as crossID and year as dateID. Now when I click on new object --> estimation and specify my equation, I get the "near singular matrix" error. I checked the correlation between my independent and dependent variables and it is "just" 30%. I'm thinking that I might make a beginners mistake when estimating. Could anyone advise me? Do I need to create a pool object first? Or is it because I have some missing values?

Many thanks

Re: Nonlinear S curve estimation with panel data

Posted: Wed Dec 15, 2021 8:10 am
by startz
Hi all,

I'm new to EViews, having used Stata for 10 years, so apologies if this is a beginners question.
I am struggling with estimating the following (S curve) equation:
Y = 1/ (C1+C2 * C3 ^X)
Nothing to do with EViews, C2 and C3 aren't separately identified. All you can get is the product.

Re: Nonlinear S curve estimation with panel data

Posted: Wed Dec 15, 2021 8:24 am
by AH2021
I see, thanks startz