CES Production Function

[Carmen Mendoza]

Hi! I want to fit a CES production function with the following variables( valueadded, kapital(capi), wages(salprod), raw materials(matprim), electricity(enelec) ) and its logarithms, I run the following command:

nl (lva={b0} -(1/{rho}*ln({alpha}*salprod^(-1*{rho})+ ({delta})*capi^(-1*{rho}))+({gammaa})*matprim^(-1*{rho}))+({sigma})*enelec^(-1*{rho})))) , but Stata gives the error message "varlist not allowed"

Another option that I´ve tried is

nl (lva={b0} -(1/{rho}*ln({delta}*salprod^(-1*{rho})+ (1-{delta})*capi^(-1*{rho})))),initial (rho 1 delta 0.5)
with only two variables, and Stata gives the error message "starting values invalid or some RHS variables have missing values"

What can I do??

Carmen

[Nick Cox]

nl (lva={b0} -(1/{rho}*ln({alpha}*salprod^(-1*{rho})+ ({delta})*capi^(-1*{rho}))+({gammaa})*matprim^(-1*{rho}))+({sigma})*enelec^(-1*{rho}))))

I saved this command text as a file and ran it through hexdump and found a problem, namely unmatched parentheses:

    040  28  (                  10
    041  29  )                  12

NB: please use CODE delimiters as requested. Commands like that are hard to read at the best of times, and harder still as ordinary text.

[Carmen Mendoza]

Yes, I correct this but now, the error message is "varlist no allowed"

[Nick Cox]

I don't even know what the function should be. Not likely to be a problem, but

-{rho}

is surely simpler than

-1*{rho}

To play with this, we would need a small data example with values of

 
 lva salprod capi matprim enelec