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

Re: st: Threshold regression

From   "Alessandro Gambini" <>
To   <>
Subject   Re: st: Threshold regression
Date   Thu, 7 May 2009 11:46:09 +0200

Dear Maarten,

Thank you very much for your kind reply and for suggesting me how to properly formulate a question in Statalist.
I will try to formulate my question more precisely.

I would like to run a firms-sample splitting on a continously-distributed variable, such as firm size. This problem in firms' growth literature is dealt in Hansen B. (2000) "Sample splitting and trheshold estimation" Econometrica vol 68 pp. 575-603 and Chan KS (1993) "Consistency and Limiting distribution of the least squares estimator of a threshold autoregressive model", The Annals of Statistics, vol 21 pp. 520-533.

I have found the Stata command "threg" to "fit threshold regression model", which at first sight could help me in doing it, but I cannot undertand well the sintax and the help does not explain what the options "lnx0", "mu", "time", "failure" represent. Furthermore, the help of "threg" does not provide any reference, so I am not sure if this the right command to solve my problem and if yes how to use those options.

Can anyone help me in this?

Thank you.


----- Original Message ----- From: "Maarten buis" <>
To: <>
Sent: Thursday, May 07, 2009 10:28 AM
Subject: Re: st: Threshold regression

--- On Thu, 7/5/09, Alessandro Gambini wrote:
I would like to run a threshold rgeression as shown in
Hansen (2000) in order to split my sample of firms in
subsamples by size. I would like to know if the Stata
"threg" command is the right one to do it.

What the people on Statalist have in common is that they
use Stata, but they tend to come from very different
disciplines. As a consequence, a reference in terms of
only an author and year of publication that is
completely obvious in your (sub-sub-)discipline is
probably completely meaningles to most members of this
list. That is why the Statalist FAQ askes everybody
to provide full references and not just an author/year
reference (a link to that FAQ is at the bottom of every
message on statalist).

Also when refering to a user written program, you are asked
to tell us where you got it from. This has two reasons, 1)
we may not be able to find that program, especially if that
user written program has not been submitted to either SSC
or the Stata Journal, and 2) there may be multiple versions
floating around in cyber space, and a problem someone
reports may already have been solved in a newer version
available from another source.

-- Maarten

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index