Forum: help
Monitor Forum | Start New ThreadRE: Imposing regional monotonicity - JPA paper 2009 [ Reply ] By: Arne Henningsen on 2017-01-07 15:52 | [forum:43783] |
1) Yes, the standard errors of *all* estimated parameters (i.e. including the z-variables) may be biased due to the use of a generated regressor. 2) No, you do not have to re-estimate step 3 again. caCoef are the final (restricted) coefficient estimates. |
Imposing regional monotonicity - JPA paper 2009 [ Reply ] By: Woon Kan Yap on 2016-12-29 18:21 | [forum:43769] |
Dear Dr Henningsen Referring to the above article, I would like to ask a few questions as below: 1) Pg 220 - as Eq.12 includes a generated regressor, it is said that the standard errors obtained in the third step might be biased. I would like to confirm would the standard errors for the z-variables be affected as well? If the standard errors are biased, would they be biased upwards? 2) Pg 220 - In regards to Eq.14, it is said that an adjustment is necessary on parameters alpha_0 and alpha_1 and the commands for the adjustment are provided in appendix. I seem to be a bit loss at this juncture. What should I do with caCoef - the adjusted coefficients of the restricted production frontier? Do I have to re-estimate step 3 again? Thanks |