MLE estimation

[Byeyo]

Hello,

I have an estimation problem while doing MLE estimation.
I think something is wrong, but I cannot find it.
My simple program is below.

logl logl1
logl1.append @logl loga
logl1.append res=lnep-c(1)-c(2)*lnecap-c(3)*lnco2-c(4)*lnl
logl1.append a1=log(@cnorm(-(res*c(7))/c(6)))
logl1.append loga = c(5)-log(c(6))+a1-1/2/(c(6)^2)*(res^2)
logl1.ml(showopts)

However, while doing it, the program sends a message
"Missing values in @logl series at current coefficients at observation 1 in "DO_LOGL.ML(SHOWOPTS)"

Please, help me!!

Thank you for your advanced help.
Best regards,

[Byeyo]

Whole values are positive, so logl values are not "NA" values..
And I try to other starting values,
c(1) to c(4) for OLS coefficient,
and c(5) for 100, and c(6) and c(7) for 1, each.

However, the result is like this


Coefficient //Std. Error//z-Statistic//Prob.

C(1) NA 1225.754 NA NA
C(2) NA 2.142668 NA NA
C(3) NA 1.580528 NA NA
C(4) NA 2.603311 NA NA
C(7) NA 4992.512 NA NA
C(6) NA 647.4918 NA NA
C(5) NA 54.93822 NA NA

[EViews Gareth]

Really does sound like bad starting values, you might need to experiment some more.