[Limdep Nlogit List] GMXL (WTP space) in Nlogit

William Greene wgreene at stern.nyu.edu
Sat Mar 7 00:19:26 AEDT 2020

Dear Thamarasi Kularatne. Unfortunately, no, it will not work.  NLOGIT's
routine is specifically written to
expect a single normalized coefficient.  The computations are specifically
programmed with a single
normalization.  What you suggest below would require the full set of
parameters for each alternative
to be normalized separately.  For those parameters that are common across
alternatives, it's not clear
how this could be achieved while keeping the model internally consistent.
Bill Greene

On Thu, Mar 5, 2020 at 7:36 PM Thamarasi Kularatne via Limdep <
limdep at mailman.sydney.edu.au> wrote:

> Hi,
> I'm trying to run a generalised mixed logit model (GMXL) in WTP space in
> order to account for taste and scale heterogeneity.
> As explained in the Nlogit 6 reference guide, we can choose one of the
> parameters in the GMXLOGIT model to have a coefficient of one and build
> this nonlinearity into the GMXLOGIT model by changing its type to (*type)
> in the ; Fcn = (*type),… specification.
> In my case, I have 3 alternatives and the cost parameters need to be
> alternative specific. Therefore when estimating GMXL in WTP space, I need
> to specify 3 fixed parameters. example below:
> GMXlogit ;userp
> ;lhs=choice,cset,altij
> ;choices=oral,inject,infuse,none
> ;pds=9 ;halton ;pts=10
> ;fcn=ocost(*n),icost(*n),ivcost(*n),eff(n),mild(n),inj_6m(n),spec(n)
> ;par
> ;pwt
> ;gmx
> ;cor
> ;gamma=0
> ;tau=0.1
> ;Model:
> U(oral)=eff*eff+ oral_wee*oral_wee + oral_m*oral_mon + mild*mild_se +
> sev2*sev_2 + ocost*oralcost+ spec*spec+ oself*oral_sel /
> U(inject)=inject+eff*eff+ inj_w*inj_week + inj_m*inj_mon + inj_6m*inj_6mon
> + inj_y*inj_yr + mild*mild_se + sev2*sev_2+ icost*injcost+
> spec*spec+iadmin*admin_in /
> U(infuse)=infuse+ eff*eff+ mild*mild_se + sev2*sev_2+ivcost*ivcost+
> iv_6m*iv_6mon + iv_yr*iv_yr + spec*spec + ivadmin*admin_iv/
> U(none)=None $
> This does not work. Nlogit only takes one parameter at a time with a * in
> the fcn specification. Is there a way for me to run the GMXL in WTP space
> with 3 alternative specific cost parameters?
> Thanks,Thames
> _______________________________________________
> Limdep site list
> Limdep at mailman.sydney.edu.au
> http://limdep.itls.usyd.edu.au

William Greene
Department of Economics, emeritus
Stern School of Business, New York University
44 West 4 St.
New York, NY, 10012
URL: https://protect-au.mimecast.com/s/77s8Ck8vAZtOnrP2Qt2XbIf?domain=people.stern.nyu.edu
Email: wgreene at stern.nyu.edu
Ph. +1.646.596.3296
Editor in Chief: Journal of Productivity Analysis
Editor in Chief: Foundations and Trends in Econometrics
Associate Editor: Economics Letters
Associate Editor: Journal of Business and Economic Statistics

More information about the Limdep mailing list