[Limdep Nlogit List] elasticities in nl models

William Greene wgreene at stern.nyu.edu
Tue Jul 29 23:58:51 EST 2008


Philomela. Nothing is wrong. The equal elasticities is a signature
feature of the MNL. The MNL model applies within the branches of a
nested logit model.  By forcing the IV parameter to equal 1.0, you
also forced the equal elasticities result across the branches as you
see in your table.
/Bill Greene

----- Original Message -----
From: Philomela at gmx.net
To: "Limdep and Nlogit Mailing List" <limdep at limdep.itls.usyd.edu.au>
Sent: Tuesday, July 29, 2008 9:31:21 AM (GMT-0500) America/New_York
Subject: Re: [Limdep Nlogit List] elasticities in nl models

Dear Bill, 
thanks very much for answering my question.  sorry for asking again about the elasticities. When I run the following model:

nlogit
;lhs= choice, cset, alti
;choices= tap,public, borehole, other
;tree=A(public,tap), B( borehole, other)
;start=logit
;ivset:(A)=[1.0]
;maxit=1000
;RU2
;model: 
U(tap) = asctap  + frequ1*frequ1 + frequ2*frequ2  +tapp*price + tsump*sump+ 
sexp*sexp + ascavai*ascavai +  ascemply*ascemply + tlincp*lincp/ 
U(public) = ascpublic + pubdist1*pubdist1 + pubdist2*pubdist2 +  frequ1*frequ1 
+ frequ2*frequ2 + wait2*wait2+ pp*price+ psump*sump+ ascavai3*ascavai3+ plincp*lincp/
U(borehole) = ascbore+  bp*price + bsump*sump 
+sexp*sexp + ascavai2*ascavai2 +  ascball*ascball+ blincp*lincp/
U(other)=0 ;
Effects: price(*)$

I get equal cross elasticities as follows:

Decomposition of Effect           Total |
      |                        Trunk   Limb   Branch   Choice     Effect|
      | Trunk=Trunk{1}                                                  
      | Limb=Lmb[1|1]                                                   
      |    Branch=A                                                     
      | *     Choice=PUBLIC     .000   .000  -2.059  -4.392     -6.450  
      |       Choice=TAP        .000   .000  -2.059   4.412      2.353  
      |    Branch=B                                                     
      |       Choice=BOREHOLE   .000   .000   2.353    .000      2.353  
      |       Choice=OTHER      .000   .000   2.353    .000      2.353  

 Decomposition of Effect           Total 
      |                        Trunk   Limb   Branch   Choice     Effect|
      | Trunk=Trunk{1}                                                  
      | Limb=Lmb[1|1]                                                   
      |    Branch=A                                                     
      |       Choice=PUBLIC     .000   .000  -3.755   7.399      3.644  
      | *     Choice=TAP        .000   .000  -3.755  -7.798    -11.553  
      |    Branch=B                                                     
      |       Choice=BOREHOLE   .000   .000   3.644    .000      3.644 
      |       Choice=OTHER      .000   .000   3.644    .000      3.644  

Decomposition of Effect           Total |
      |                        Trunk   Limb   Branch   Choice     Effect|
      | Trunk=Trunk{1}                                                  
      | Limb=Lmb[1|1]                                                   
      |    Branch=A                                                     
      |       Choice=PUBLIC     .000   .000   3.257    .000      3.257  
      |       Choice=TAP        .000   .000   3.257    .000      3.257  
      |    Branch=B                                                     
      | *     Choice=BOREHOLE   .000   .000  -3.511  -1.544     -5.055  
      |       Choice=OTHER      .000   .000  -3.511  13.268      9.757  
     
So for tap and public i get equal cross elasticities. What is wrong?

Thanks so much,Philo


-------- Original-Nachricht --------
> Datum: Mon, 28 Jul 2008 11:58:10 -0400 (EDT)
> Von: William Greene <wgreene at stern.nyu.edu>
> An: Limdep and Nlogit Mailing List <limdep at limdep.itls.usyd.edu.au>
> Betreff: Re: [Limdep Nlogit List] elasticities in nl models

> Philo. The ;Effects: routine in NLOGIT works for all model specifications.
> With respect to a nested logit model, however, keep in mind that you can
> only
> compute effects of attributes on elemental alternatives. Nothing that
> appears
> in any utility functions for branches or higher can be analyzed.
> /Bill Greene
> 
> ----- Original Message -----
> From: Philomela at gmx.net
> To: limdep at limdep.itls.usyd.edu.au
> Sent: Monday, July 28, 2008 9:33:48 AM (GMT-0500) America/New_York
> Subject: [Limdep Nlogit List] elasticities in nl models
> 
> Dear all, 
> 
> i want to estimate direct and cross elasticties for a nested logit model.
> Can I use the ;effects command for that or does it need to be modified to
> account for the correlation of alternatives within the same partition?
> 
> Thanks very much, 
> Philo
> -- 
> GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen!
> Jetzt dabei sein:
> http://www.shortview.de/wasistshortview.php?mc=sv_ext_mf@gmx
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au
> 
> 
> -- 
> Professor William Greene
> Department of Economics
> Stern School of Business
> New York University
> 44 West 4th St., Rm. 7-78
> New York, NY   10012
> http://www.stern.nyu.edu/~wgreene
> 212.998.0876
> 
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au

-- 
Psssst! Schon das coole Video vom GMX MultiMessenger gesehen?
Der Eine für Alle: http://www.gmx.net/de/go/messenger03
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au


-- 
Professor William Greene
Department of Economics
Stern School of Business
New York University
44 West 4th St., Rm. 7-78
New York, NY   10012
http://www.stern.nyu.edu/~wgreene
212.998.0876




More information about the Limdep mailing list