nonlinear panel estimation

[mikenike]

I am trying to do a simple nonlinear estimation and I get many errors, mostly "insufficient number of observations". Moreover, sometimes I am trying to re-estimate a previous estimation and I get "Near singular matrix). It may be my machine, but I am attaching the workfile, just in case somebody wants to try. the equation I am shooting for is nbd=500/(1+C(1)*exp(-C(2)gdpc1000)). The last message i got is "NAGDPC1000 is not defined".

I wasted the whole morning, help !

[EViews Gareth]

You're missing a "*" between the coefficient and gdpc1000,


With regards to the insufficient observations, and changing results, they're probably down to what you have as starting values - pay attention to what is in C before you estimate.

[mikenike]

How do I specify starting values??? I did not see any dialog.

[mikenike]

You're missing a "*" between the coefficient and gdpc1000,


With regards to the insufficient observations, and changing results, they're probably down to what you have as starting values - pay attention to what is in C before you estimate.

How do I specify starting values??? I did not see any dialog. I open an equation and I input the nonlinear specification. What else should I do. I cannot find any examples/help anywhere. And C only has C1...

[EViews Gareth]

http://bfy.tw/BtiE