Estimation of Restricted CobbDouglas Production Function
Moderators: EViews Gareth, EViews Moderator
Estimation of Restricted CobbDouglas Production Function
How I can estimate the restricted CobbDouglas Production Function with three factor inputs of the form:
Y = A*(K^alpha)(L^beta)(N^gamma); alpha + beta + gamma = 1
where
Y = Output
K = Capital
L = Labor, and
N = Land
Please help me out with the estimating of this function in Eviews.
Y = A*(K^alpha)(L^beta)(N^gamma); alpha + beta + gamma = 1
where
Y = Output
K = Capital
L = Labor, and
N = Land
Please help me out with the estimating of this function in Eviews.

 Nonnormality and collinearity are NOT problems!
 Posts: 3466
 Joined: Wed Sep 17, 2008 2:25 pm
Re: Estimation of Restricted CobbDouglas Production Function
You can take logs and do a linear regression or you can do a nonlinear regression.
Perhaps you should tell what us you've done so far and where you got stuck.
Perhaps you should tell what us you've done so far and where you got stuck.
Re: Estimation of Restricted CobbDouglas Production Function
Thanks for your comments.
Given function:
Y = A*(K^alpha)*(N^beta)*(L^gemma)
where,
K = Capital
N = Labour
L = Land
and, given restriction:
alpha + beta + gemma = 1
I want to run a regression of the form.
log(Y) = A + (1  beta  gemma)*log(K) + (1alpha  gemma)*log(N) + (1alpha  beta)*log(L)
I failed to run this regression in Eviews.
Please help me either with the estimation of the function in Eviews and with the functional form.
Thanks.
Given function:
Y = A*(K^alpha)*(N^beta)*(L^gemma)
where,
K = Capital
N = Labour
L = Land
and, given restriction:
alpha + beta + gemma = 1
I want to run a regression of the form.
log(Y) = A + (1  beta  gemma)*log(K) + (1alpha  gemma)*log(N) + (1alpha  beta)*log(L)
I failed to run this regression in Eviews.
Please help me either with the estimation of the function in Eviews and with the functional form.
Thanks.

 Nonnormality and collinearity are NOT problems!
 Posts: 3466
 Joined: Wed Sep 17, 2008 2:25 pm
Re: Estimation of Restricted CobbDouglas Production Function
Imran Ali wrote:Thanks for your comments.
Given function:
Y = A*(K^alpha)*(N^beta)*(L^gemma)
where,
K = Capital
N = Labour
L = Land
and, given restriction:
alpha + beta + gemma = 1
I want to run a regression of the form.
log(Y) = A + (1  beta  gemma)*log(K) + (1alpha  gemma)*log(N) + (1alpha  beta)*log(L)
I failed to run this regression in Eviews.
Please help me either with the estimation of the function in Eviews and with the functional form.
Thanks.
In EViews, coefficients are named c(1), c(2), etc., rather than "beta,"
so you want to do something like
log(y) = c(1) +c(2)*log(k) + c(3)*log(N) + (1c(2)c(3))*log(L)
Re: Estimation of Restricted CobbDouglas Production Function
Once again thanks for your help. But my this does not solve my problem. See if i run this regression i did not get the sum of all three coefficients equal to one. So please tell me if there's any problem with the functional form. How do I specify the CobbDouglass Production function with three inputs and neutral technology in the Eviews. Regards

 Nonnormality and collinearity are NOT problems!
 Posts: 3466
 Joined: Wed Sep 17, 2008 2:25 pm
Re: Estimation of Restricted CobbDouglas Production Function
Maybe you should post exactly what you are doing, so people can take a look.
Re: Estimation of Restricted CobbDouglas Production Function
The discussion under the Putting limits on estimated coefficient values thread might help. Please always do a simple search before posting your questions, since they may have already been answered somewhere in the forum.
Re: Estimation of Restricted CobbDouglas Production Functio
HI,
i want to estimate a translog cost function and I need to put some restrictions, I used the following model
log(y) = c(1) +c(2)*log(k) + c(3)*log(N) + (1c(2)c(3))*log(L)
I get the cofficient for c(1), c(2), c(3)
For (1c(2)c(3))*log(L) am I suppose to just do 1 minus the results for the two coefficient?
i want to estimate a translog cost function and I need to put some restrictions, I used the following model
log(y) = c(1) +c(2)*log(k) + c(3)*log(N) + (1c(2)c(3))*log(L)
I get the cofficient for c(1), c(2), c(3)
For (1c(2)c(3))*log(L) am I suppose to just do 1 minus the results for the two coefficient?

 Nonnormality and collinearity are NOT problems!
 Posts: 3466
 Joined: Wed Sep 17, 2008 2:25 pm
Re: Estimation of Restricted CobbDouglas Production Functio
Those don't look like the right restrictions for a CobbDouglas.
Re: Estimation of Restricted CobbDouglas Production Functio
this si the translog cost function
ln(C(q;wL,wK,wM)) = c + cq * ln(q) + cL * ln(wL) + cK * ln(wK) + cM * log(wM)
+ .5 * [dqq * ln(q)^2 + dLL * ln(wL)^2 + dKK * ln(wK)^2 + dMM * ln(wM)^2]
+ .5 * [(dLK + dKL) * ln(wL)*ln(wK) + (dLM + dML) * ln(wL)*ln(wM) + (dKM + dMK) * ln(wK)*log(wM)]
+ dLq * ln(wL)*ln(q) + dKq * ln(wK)*ln(q) + dMq * ln(wM)*ln(q)
and the restrictions are:
1 = cL + cK + cM
0 = dLL + dLK + dLM
0 = dKL + dKK + dKM
0 = dML + dMK + dMM
0 = dLq + dKq + dMq
0 = dLq, 0 = dKq, 0 = dMq, 1 = cq
ln(C(q;wL,wK,wM)) = c + cq * ln(q) + cL * ln(wL) + cK * ln(wK) + cM * log(wM)
+ .5 * [dqq * ln(q)^2 + dLL * ln(wL)^2 + dKK * ln(wK)^2 + dMM * ln(wM)^2]
+ .5 * [(dLK + dKL) * ln(wL)*ln(wK) + (dLM + dML) * ln(wL)*ln(wM) + (dKM + dMK) * ln(wK)*log(wM)]
+ dLq * ln(wL)*ln(q) + dKq * ln(wK)*ln(q) + dMq * ln(wM)*ln(q)
and the restrictions are:
1 = cL + cK + cM
0 = dLL + dLK + dLM
0 = dKL + dKK + dKM
0 = dML + dMK + dMM
0 = dLq + dKq + dMq
0 = dLq, 0 = dKq, 0 = dMq, 1 = cq

 Nonnormality and collinearity are NOT problems!
 Posts: 3466
 Joined: Wed Sep 17, 2008 2:25 pm
Re: Estimation of Restricted CobbDouglas Production Functio
Then yes, "c(4)" = 1 c(2) c(3)
Re: Estimation of Restricted CobbDouglas Production Functio
Hello,
I am running a regression with a linear restriction on a coefficient as in
Y = alpha*X1 + beta*X2 + (1alphabeta)*X3 + error
Eviews only returns the estimates for alpha and beta as well as their standard errors.
To get the estimate of the coefficient for X3, I do manually 1alphabeta.
However, how do I compute the standard error of this paramater?
Thanks
I am running a regression with a linear restriction on a coefficient as in
Y = alpha*X1 + beta*X2 + (1alphabeta)*X3 + error
Eviews only returns the estimates for alpha and beta as well as their standard errors.
To get the estimate of the coefficient for X3, I do manually 1alphabeta.
However, how do I compute the standard error of this paramater?
Thanks
Re: Estimation of Restricted CobbDouglas Production Function
Hello Startz,
Many thanks indeed for your valuable help.
Much appreciated.
Kind regards.
Who is online
Users browsing this forum: No registered users and 5 guests