Forum: help

RE: Combination of CES and Cobb-Douglas- micEconCES [ Reply ]
By: Esther delesalle on 2015-02-27 11:05

[forum:41951]

Dear Arne,

I thank you a lot for your quick answer.

That's what I was afraid of. Ok, I will check the nls function and otherwise try to extend the code of the micEconces package.

I have already constrained the value of delta1 (CES with 4 inputs) between 0 and 1 but results are not very economically meaningful (delta1=1 which has no sense in my case). Besides, according to the optimisation method, the value of rho1 seems very sensible to the value of delta1. This is why I would like to fix the value of the parameter to the value observed in my data set. I just remembered that I can reduce the interval of values thanks to the upper and lower arguments.

I would like also to ask you a last question if I may :
For about 15% of my observations, the output is null and I cannot run a Kmenta approximation without adding +1.
Then, when I run a CES with my ouptut without +1, I wonder if the command cesCalc succeeds to compute the ouput with the Kmenta approximation when rho=0. I don't have any message error and I don't know if I have to add 1 to my ouptut.

I also take this opportunity to tell you that your paper « Econometric Estimation of the « Constant Elasiticity of Substitution » Function in R : package micEconCES » is very useful and very interesting.

Best regards.
Esther

RE: Combination of CES and Cobb-Douglas- micEconCES [ Reply ]
By: Arne Henningsen on 2015-02-26 20:36

[forum:41950]

Dear Esther

The micEconCES package cannot estimate the model specification that you describe (combination of a CES function and a Cobb-Douglas function. You may be able to estimate this functional forrm with function nls(). Otherwise you could extend the code of the micEconCES package to include this model specification.

How do you want to constrain the value of delta? To be between zero and one?

Best regards,
Arne

Combination of CES and Cobb-Douglas- micEconCES [ Reply ]
By: Esther delesalle on 2015-02-26 17:51

[forum:41949]

Dear all,

I am trying to estimate a CES function with the micEconCES package. I consider two inputs x1 and x2 in the CES function and I would like to introduce other inputs which have a Cobb-Douglas relationship with my CES inputs. That is to say, I would like to express:
y = gamma * ( delta * x1^(-rho) + ( 1 - delta ) * x2^(-rho) )^(-nu/rho)
with gamma which represents all other inputs:
gamma= x3^alpha3.x4^alpha4....xn^alphan when I have n additional inputs.

I have two questions:
1) does someone know if it is possible? In this case, how can I define gamma?
2) Do you know how I can constrain the value of delta?

I am starting to learn R and I would be very grateful if someone could answer me.
Best.
Esther