Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Nested logit, bivariate probit or Heckman probit?
From
Sriram Narayanamoorthy <[email protected]>
To
[email protected]
Subject
Re: st: Nested logit, bivariate probit or Heckman probit?
Date
Mon, 26 Aug 2013 17:49:52 -0700
I am not familiar with estimating Nested logit models in Stata - but
by looking at your specification - there is an identification issue.
If you are specifying a generic variable like "firm size" in this case
- you will have to normalize the nesting parameter. So essentially you
will have to set the nesting parameter (theta) for you top nest to be
same and estimate it as a free parameter (the parameter should be less
than 1). Such models are called normalized nested logit models.
On Mon, Aug 26, 2013 at 5:21 PM, XIAOFEI XING <[email protected]> wrote:
> Dear Statalist,
>
> I investigate in whether big firms tend to retain top-tier financial advisors to make takeover deals and whether they tend to make in-house deals (no financial advisors retained). I specify a two-level nested model. The tree is shown as follows:
>
> Advisor Selection
> Retaining Advisors In-House Deals
> Top-Tier Advisors Non-Top-Tier Advisors In-House Deals
>
>
> Level 1 is retaining advisors/in-house deals contrast. In other words, I investigate whether big firms tend to retain financial advisors or conduct in-house deals.
>
> In Level 2, for deals that retain advisors, I investigate whether big firms tend to retain top-tier advisors rather than non-top-tier advisors.
>
> Therefore, in each level, I should include a same variable - "FirmSize". However, Stata tell me: "variable FirmSize is specified in more than one equation; this is not allowed."
>
> How can I address this issue? Does it mean that the same independent variable cannot be used in the two levels of a nested logit model?
>
> I also check the definition of bivariate probit model and Heckman probit. However, I don't think my case is suitable to the bivariate probit or Heckman probit.
>
> Thanks a lot.
>
> Xiaofei Xing
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
--
Sriram Narayanamoorthy
Graduate Research Assistant
email: [email protected]
Department of Civil, Architectural and Environmental Engineering
The University of Texas at Austin
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/