[Limdep Nlogit List] how to replicate the WTP_I values in RPL model?
sabouha2 at gmail.com
Fri Mar 27 22:37:31 EST 2009
I had done something similar using RPM and triangular and normal
distribution. The way I did mine is my attributes A, B were revealed as
random but made sure the cost is fixed just for the ease of computation.
There are references that have used cost as fixed.
On Mon, Mar 16, 2009 at 4:18 PM, <scolombo at ugr.es> wrote:
> Dear all,
> I am estimating individual WTP in a RPL model. I used the command WTP
> =attribute a / cost. Both attributes are specified as random: attribute A
> is normal, cost is triangular)
> I observed that the individual WTP is not the simple division of the
> individual coefficient of attribute A and cost. In the manual it is
> explained that the coefficient and standard deviation of each individual
> is the mean and standard deviation of the conditional distribution of each
> individual (that is , the distribution obtained by considering all the
> people who did the same choices (I do not allow for heterogeneity in the
> mean, neither in the variance of the error)).
> I would like to replicate (or at least get very close)the WTP values of
> the matrix WTP_I.
> My question is: Do I obtain the same individual WTP if I average the
> ratio between a draw from the normal distribution of attribute A (I have
> for each individual his own beta and standard deviation) and a draw from
> the tringular distribution of the cost attribute?
> I hope that it is clear what I woulld like to do. Did anybody do something
> Many thanks in advance for any advises.
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
PhD Student (Economics)/Part-time Academic & Resident Tutor
Dept. of Economics & Int'l Development
e-mail: s.abdullah at bath.ac.uk
More information about the Limdep