|
Activity Number:
|
445
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics and Marketing
|
| Abstract - #305261 |
|
Title:
|
Estimating Willingness To Pay with Random Coefficient Models
|
|
Author(s):
|
Garrett Sonnier*+
|
|
Companies:
|
University of California, Los Angeles
|
|
Address:
|
110 Westwood Plaza, Los Angeles, CA, 90095,
|
|
Keywords:
|
choice ; logit ; heterogeneity
|
|
Abstract:
|
It is well-known that the choice probabilities derived from random utility models suffer from scale invariance. Only ratios of parameters can be identified, and the usual approach to this identification problem is to place some type of restriction on the parameter space. Generally, the specific restriction employed has been regarded as an arbitrary choice. In models with homogeneous coefficients or latent-class approaches to heterogeneity, the model fits are unaffected and point estimates of parameters obtained under one type of restriction can be recovered easily by transforming those from another restriction. This typically leads researchers to choose the parameterization that most easily facilitates estimation. We will demonstrate that the choice of parameterization in a random coefficient setting incorporating parametric distributions of heterogeneity matters.
|
