Estimation of Binary Response Models With Endogenous Regressors

Changhui Kang, Myoung-jae Lee

Research output: Contribution to journalArticle

Abstract

This paper reviews six approaches to binary response (y<inf>1</inf>) structural forms with an endogenous regressor y<inf>2</inf>: (i) the two-stage least squares estimator-like substitution approach, (ii) the control function approach, (iii) the system reduced-form approach, (iv) the artificial instrumental regressor approach, (v) the transformed-response instrumental variable estimator approach and (vi) the classical maximum likelihood estimator approach. The applicability of the six methods differs greatly, depending on whether y<inf>2</inf> is a continuously distributed random variable or a discrete transformation of a latent y 2 *. We conduct a real-data-based simulation study, and provide an empirical illustration. Our overall recommendation is using (i) and (ii), as the others have undesirable features such as analytic complexity in (iii), computational difficulty in (iv) and (vi), and poor finite-sample performance in (v).

Original languageEnglish
Pages (from-to)502-530
Number of pages29
JournalPacific Economic Review
Volume19
Issue number4
DOIs
Publication statusPublished - 2014 Oct 1

Fingerprint

Random variables
substitution
Maximum likelihood
Substitution reactions
simulation
performance
structural response

ASJC Scopus subject areas

  • Geography, Planning and Development
  • Development
  • Aerospace Engineering

Cite this

Estimation of Binary Response Models With Endogenous Regressors. / Kang, Changhui; Lee, Myoung-jae.

In: Pacific Economic Review, Vol. 19, No. 4, 01.10.2014, p. 502-530.

Research output: Contribution to journalArticle

@article{9881cc48fbfe42619226998eb287023f,
title = "Estimation of Binary Response Models With Endogenous Regressors",
abstract = "This paper reviews six approaches to binary response (y1) structural forms with an endogenous regressor y2: (i) the two-stage least squares estimator-like substitution approach, (ii) the control function approach, (iii) the system reduced-form approach, (iv) the artificial instrumental regressor approach, (v) the transformed-response instrumental variable estimator approach and (vi) the classical maximum likelihood estimator approach. The applicability of the six methods differs greatly, depending on whether y2 is a continuously distributed random variable or a discrete transformation of a latent y 2 *. We conduct a real-data-based simulation study, and provide an empirical illustration. Our overall recommendation is using (i) and (ii), as the others have undesirable features such as analytic complexity in (iii), computational difficulty in (iv) and (vi), and poor finite-sample performance in (v).",
author = "Changhui Kang and Myoung-jae Lee",
year = "2014",
month = "10",
day = "1",
doi = "10.1111/1468-0106.12076",
language = "English",
volume = "19",
pages = "502--530",
journal = "Pacific Economic Review",
issn = "1361-374X",
publisher = "Wiley-Blackwell",
number = "4",

}

TY - JOUR

T1 - Estimation of Binary Response Models With Endogenous Regressors

AU - Kang, Changhui

AU - Lee, Myoung-jae

PY - 2014/10/1

Y1 - 2014/10/1

N2 - This paper reviews six approaches to binary response (y1) structural forms with an endogenous regressor y2: (i) the two-stage least squares estimator-like substitution approach, (ii) the control function approach, (iii) the system reduced-form approach, (iv) the artificial instrumental regressor approach, (v) the transformed-response instrumental variable estimator approach and (vi) the classical maximum likelihood estimator approach. The applicability of the six methods differs greatly, depending on whether y2 is a continuously distributed random variable or a discrete transformation of a latent y 2 *. We conduct a real-data-based simulation study, and provide an empirical illustration. Our overall recommendation is using (i) and (ii), as the others have undesirable features such as analytic complexity in (iii), computational difficulty in (iv) and (vi), and poor finite-sample performance in (v).

AB - This paper reviews six approaches to binary response (y1) structural forms with an endogenous regressor y2: (i) the two-stage least squares estimator-like substitution approach, (ii) the control function approach, (iii) the system reduced-form approach, (iv) the artificial instrumental regressor approach, (v) the transformed-response instrumental variable estimator approach and (vi) the classical maximum likelihood estimator approach. The applicability of the six methods differs greatly, depending on whether y2 is a continuously distributed random variable or a discrete transformation of a latent y 2 *. We conduct a real-data-based simulation study, and provide an empirical illustration. Our overall recommendation is using (i) and (ii), as the others have undesirable features such as analytic complexity in (iii), computational difficulty in (iv) and (vi), and poor finite-sample performance in (v).

UR - http://www.scopus.com/inward/record.url?scp=84939227153&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939227153&partnerID=8YFLogxK

U2 - 10.1111/1468-0106.12076

DO - 10.1111/1468-0106.12076

M3 - Article

VL - 19

SP - 502

EP - 530

JO - Pacific Economic Review

JF - Pacific Economic Review

SN - 1361-374X

IS - 4

ER -