I am working on translog production function. I tried to estimate it through share equations but log linear share equations were unable to predict well. Then I introduced a square term of the factor input in the share equation (i.e. sk(share of capital) was only added with square of capital in addition to linear equation). 

As Translog function has properties like sum of shares is 1 and its estimation require to one equation to drop to avoid singularity. The result are invariant to which equation is dropped, but my results vary.

How to solve this? How can I expand translog production fn as cubic function??

A file of results is attached

Please reply

More Sajid Hussain's questions See All
Similar questions and discussions