Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Decompose the technical efficiency

From   Scott Merryman <>
Subject   Re: st: Decompose the technical efficiency
Date   Sat, 11 Jun 2011 21:14:56 -0500

Isn't necessary to calculate the relative inefficiency (r(max) - u) first?

For example:

webuse xtfrontier1, clear
xtreg lnwid lnwo lnmac,re
predict u,u
sum u
gen te_re = exp(-(r(max) - u))
sum te


On Sat, Jun 11, 2011 at 6:03 AM, Song Nhac <> wrote:
> Dear All,
> I employ the command as:
> xtreg profit loan earningasset priceoflabor priceoffund equity totalasset, re
> to find the technical efficiency of banks. And decompose the effect
> u_i as the inefficiency and TE=exp(-u_i) ,
> I know that in this command, we assume that ui~Gaussian. But I found
> that the TE is greater than 1. Anyone know how I put more assumption
> to u_i? Since I run xtfrontier...., ti- it gave me a quite different
> TE.
> My purpose is to compare the unobservable part with the TE, so that
> why I intend to employ xtreg since from xtfrontier- I cannot decompose
> the error term or get the value of v_it.
> Thank for your kindly attention and support.
> Great weekend.
> Song Nhac
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index