Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Scott Merryman <scott.merryman@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: inverse logit of gamma and xtfrontier |
Date | Fri, 17 Feb 2012 12:34:58 -0600 |
In the -xtfrontier- command (stochastic frontier models for panel data), the gamma parameter is parameterized in terms of the inverse logit of gamma and Stata reports the hypothesis test that the inverse logit of gamma is 0. Does this reported hypothesis test has any substantive meaning? It not clear to me what inferences one can draw if the inverse logit of gamma is not statistically different from 0 (and therefore gamma is 0.50). To test for existence of inefficiency effects, following Battese and Coelli (1992), one would need to constrain the three parameters mu = eta = gamma = 0 and use a likelihood ratio test: webuse xtfrontier1 constraint 1 [eta]_cons = 0 constraint 2 [mu]_cons = 0 constraint 3 [ilgtgamma]_cons = -100 /* making gamma close to zero */ xtfrontier lnwidgets lnmachines lnworkers, tvd est store A xtfrontier lnwidgets lnmachines lnworkers, tvd constraints(1 2 3) lrtest A ., stat disp (1 - chi2(1, r(chi2)))*.5 Thank you, Scott Battese and Coelli. 1992 "Frontier production functions, technical efficiency and panel data: With application to paddy farmers" Journal of Productivity Analysis,Volume 3, Numbers 1-2. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/