Forum: open-discussion
Monitor Forum | 
Start New Thread
| RE: efficiencies() [ Reply ] By: Arne Henningsen on 2015-05-10 19:48 | [forum:42221] |
|
Dear João The efficiencies() method calculates E[e^(-u_it)] or E[e^u_it] depending on argument "minusU". In case of a standard Cobb-Douglas or Translog production function, E[e^(-u_it)] is equal to the *expectation* of the ratio between the observed output quantity and the maximum possible ("frontier") output quantity. Observation-specific noise terms (v_it) are not known, but you could derive the equation for calculating their expectations E[v_it] and then calculate the expectations manually. Best regards, Arne |
|
| efficiencies() [ Reply ] By: João Silva on 2015-04-27 16:57 | [forum:42204] |
|
Good afternoon. I have been working with the frontier package recently in some work I am doing for my PhD. I estimated different stochastic frontier models and my question refers to the computation of the technical efficiency scores and the quantification of the statistical noise for each observation in my dataset. If I understand correctly the function efficiencies is estimating technical efficiency as the ratio between observed output/frontier output. Is this right? My other question relates to the statistical noise (v_it). How exactly can I estimate it for each observation? Thank you, João |
|
