Estimation of the preference heterogeneity within stated choice data using semiparametric varying-coefficient methods

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

This study proposes the use of semiparametric varying-coefficient methods to estimate the preference heterogeneity within stated choice data. Semiparametric varying-coefficient methods have the potential to overcome the disadvantages of conventional random parameter models and latent class models. For binary probit models with varying coefficients, in particular, this study proposes an easy-to-compute local iterative least squares (LILS) approach, based on the expectation-maximization algorithm. The finite sample properties of the LILS estimator are assessed using Monte Carlo experiments. In order to demonstrate the practical usefulness of semiparametric varying-coefficient methods, we present an empirical study, conducting an economic valuation of a landscape with dichotomous choice contingent valuations.

Original languageEnglish
Pages (from-to)1129-1148
Number of pages20
JournalEmpirical Economics
Volume45
Issue number3
DOIs
Publication statusPublished - 2013 Dec
Externally publishedYes

Fingerprint

Varying Coefficients
Valuation
Latent Class Model
Probit Model
Random Parameters
Monte Carlo Experiment
Least Squares Estimator
Expectation-maximization Algorithm
Empirical Study
Least Squares
experiment
Economics
Binary
economics
Coefficients
Preference heterogeneity
Stated choice
Estimate
Demonstrate

Keywords

  • Dichotomous-choice contingent valuation
  • Discrete choice models
  • EM algorithm
  • Preference heterogeneity
  • Stated choice data
  • Varying-coefficient models

ASJC Scopus subject areas

  • Economics and Econometrics
  • Social Sciences (miscellaneous)
  • Mathematics (miscellaneous)
  • Statistics and Probability

Cite this

Estimation of the preference heterogeneity within stated choice data using semiparametric varying-coefficient methods. / Hoshino, Tadao.

In: Empirical Economics, Vol. 45, No. 3, 12.2013, p. 1129-1148.

Research output: Contribution to journalArticle

@article{7794e6c1c5244a03ac7e7917309757f6,
title = "Estimation of the preference heterogeneity within stated choice data using semiparametric varying-coefficient methods",
abstract = "This study proposes the use of semiparametric varying-coefficient methods to estimate the preference heterogeneity within stated choice data. Semiparametric varying-coefficient methods have the potential to overcome the disadvantages of conventional random parameter models and latent class models. For binary probit models with varying coefficients, in particular, this study proposes an easy-to-compute local iterative least squares (LILS) approach, based on the expectation-maximization algorithm. The finite sample properties of the LILS estimator are assessed using Monte Carlo experiments. In order to demonstrate the practical usefulness of semiparametric varying-coefficient methods, we present an empirical study, conducting an economic valuation of a landscape with dichotomous choice contingent valuations.",
keywords = "Dichotomous-choice contingent valuation, Discrete choice models, EM algorithm, Preference heterogeneity, Stated choice data, Varying-coefficient models",
author = "Tadao Hoshino",
year = "2013",
month = "12",
doi = "10.1007/s00181-012-0646-5",
language = "English",
volume = "45",
pages = "1129--1148",
journal = "Empirical Economics",
issn = "0377-7332",
publisher = "Physica-Verlag",
number = "3",

}

TY - JOUR

T1 - Estimation of the preference heterogeneity within stated choice data using semiparametric varying-coefficient methods

AU - Hoshino, Tadao

PY - 2013/12

Y1 - 2013/12

N2 - This study proposes the use of semiparametric varying-coefficient methods to estimate the preference heterogeneity within stated choice data. Semiparametric varying-coefficient methods have the potential to overcome the disadvantages of conventional random parameter models and latent class models. For binary probit models with varying coefficients, in particular, this study proposes an easy-to-compute local iterative least squares (LILS) approach, based on the expectation-maximization algorithm. The finite sample properties of the LILS estimator are assessed using Monte Carlo experiments. In order to demonstrate the practical usefulness of semiparametric varying-coefficient methods, we present an empirical study, conducting an economic valuation of a landscape with dichotomous choice contingent valuations.

AB - This study proposes the use of semiparametric varying-coefficient methods to estimate the preference heterogeneity within stated choice data. Semiparametric varying-coefficient methods have the potential to overcome the disadvantages of conventional random parameter models and latent class models. For binary probit models with varying coefficients, in particular, this study proposes an easy-to-compute local iterative least squares (LILS) approach, based on the expectation-maximization algorithm. The finite sample properties of the LILS estimator are assessed using Monte Carlo experiments. In order to demonstrate the practical usefulness of semiparametric varying-coefficient methods, we present an empirical study, conducting an economic valuation of a landscape with dichotomous choice contingent valuations.

KW - Dichotomous-choice contingent valuation

KW - Discrete choice models

KW - EM algorithm

KW - Preference heterogeneity

KW - Stated choice data

KW - Varying-coefficient models

UR - http://www.scopus.com/inward/record.url?scp=84888433864&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84888433864&partnerID=8YFLogxK

U2 - 10.1007/s00181-012-0646-5

DO - 10.1007/s00181-012-0646-5

M3 - Article

AN - SCOPUS:84888433864

VL - 45

SP - 1129

EP - 1148

JO - Empirical Economics

JF - Empirical Economics

SN - 0377-7332

IS - 3

ER -