Page 1 of 1

I receive the error of "no valid observations"

Posted: Mon Jan 04, 2016 7:13 am
by 309blank
I am trying to estiamte a system of nonlinear equations and I am new to Eviews but have programming background. I managed to import my data and create all the necessary variables and I defined my equation system thus:

Code: Select all

param sigma .5 psi 1 alphal 0.05 alphae 0.01 ly = log(psi(1)) + sigma(1)/(sigma(1) -1) *log(kl*(exp(alphal(1)*beta*t)*(lbar^beta)*(kbar^(1-beta)))^((sigma(1) -1)/sigma(1))+(1-kl)*(exp(alphae(1)*t)*quality*ebar)^((sigma(1) -1)/sigma(1))) les = (sigma(1) -1)/sigma(1) * (log(psi(1)) + log(quality)+alphae(1)*t)+ (1-sigma(1))/sigma(1) * ly + (1+sigma(1))*log(ebar) + log(1-kl) lls = (sigma(1) -1)/sigma(1) * (log(psi(1)) + alphal(1)*beta*t + beta*log(lbar) + (1-beta) * log(kbar)) + (1-sigma(1))/sigma(1) * ly + log(kl*beta)
When I try to estimate it, it says I don't have any valid observations in my system, which is puzzling to me. I have uploaded the working file. I am a bit unfamiliar with EViews as a work environment so I feel a bit lost.

Re: I receive the error of "no valid observations"

Posted: Mon Jan 04, 2016 8:52 am
by EViews Gareth
Probably taking logs of negative numbers, resulting in too many NAs.

Re: I receive the error of "no valid observations"

Posted: Tue Jan 05, 2016 6:24 am
by 309blank
It works when I change the initial parameter specification to the following:

Code: Select all

param sigma(1) .5 psi(1) 1 alphal(1) 0.05 alphae(1) 0.01
It always takes a while to get used to how another language wants its t's crossed and its i's dotted it seems :)