Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Variance-Covariance matrix using -lincom-

From   "Stas Kolenikov" <>
Subject   Re: st: Variance-Covariance matrix using -lincom-
Date   Fri, 18 Aug 2006 13:54:53 -0500

Besides -lincom-, you can use -nlcom- to compute non-linear functions
of the estimated parameters, and you will get the delta method
standard errors along with the point estimates. If you need your B3
parameter really badly for your formulae, you can consider localizing

local B3 (1-1 - [b1]_cons - [b2]_cons)

so that then you can use something like

nlcom (alpha13: ln( [b1]_cons/`B3' ) ) (alpha23: ln ( [b2]_cons/`B3') )

On 8/18/06, JEFFREY ROSEN <> wrote:

I've estimated a three equation LES model using an MLE. Due to constraints on
the system, only two equations are estimated and the third equation has  a linear
relationship with the other two.

B1 and B2 would be estimated and B3 = 1 - B1 - B2.

I've estimated B3 using the -lincom- command:

lincom 1 - [b1]_cons - [b2]_cons

I would like to calculate the standard error of some elasicities using the delta method. In order to do so, I need the variance-covariance matrix of all of the estimates. When I review the matrix using either -estat vce- or matrix list e(V), the matrix does not include the variable B3 that was estimated using -lincom-.

Is it possible to get the matrix to include B3?

Thank you in advance,

Jeff Rosen
Graduate Student
The Ohio State University

*   For searches and help try:

Stas Kolenikov
*   For searches and help try:

© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index