A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions

A.S. Kalwij

Research output: Working paperDiscussion paperOther research output

242 Downloads (Pure)

Abstract

This study develops a two-step estimator for a panel data Tobit model based on taking first-differences of the equation of interest, under conditional mean independence assumptions.The necessary correction terms are non-standard and a substantial part is therefore devoted to the formal derivation of these correction terms.The main advantage of this estimator is that it yields estimates that are far less sensitivity to misspecification of the conditional mean independence assumption than an estimation procedure set up in levels.Monte Carlo simulations are provided in support of this.
Original languageEnglish
Place of PublicationTilburg
PublisherEconometrics
Number of pages19
Volume2004-67
Publication statusPublished - 2004

Publication series

NameCentER Discussion Paper
Volume2004-67

Fingerprint

Tobit model
Estimator
Panel data
Misspecification
Monte Carlo simulation
Two-step estimator

Keywords

  • panel data
  • estimating
  • monte carlo technique

Cite this

Kalwij, A. S. (2004). A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions. (CentER Discussion Paper; Vol. 2004-67). Tilburg: Econometrics.
@techreport{473a102b78d14add847f5ac2f2ab8c3a,
title = "A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions",
abstract = "This study develops a two-step estimator for a panel data Tobit model based on taking first-differences of the equation of interest, under conditional mean independence assumptions.The necessary correction terms are non-standard and a substantial part is therefore devoted to the formal derivation of these correction terms.The main advantage of this estimator is that it yields estimates that are far less sensitivity to misspecification of the conditional mean independence assumption than an estimation procedure set up in levels.Monte Carlo simulations are provided in support of this.",
keywords = "panel data, estimating, monte carlo technique",
author = "A.S. Kalwij",
note = "Pagination: 19",
year = "2004",
language = "English",
volume = "2004-67",
series = "CentER Discussion Paper",
publisher = "Econometrics",
type = "WorkingPaper",
institution = "Econometrics",

}

Kalwij, AS 2004 'A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions' CentER Discussion Paper, vol. 2004-67, Econometrics, Tilburg.

A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions. / Kalwij, A.S.

Tilburg : Econometrics, 2004. (CentER Discussion Paper; Vol. 2004-67).

Research output: Working paperDiscussion paperOther research output

TY - UNPB

T1 - A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions

AU - Kalwij, A.S.

N1 - Pagination: 19

PY - 2004

Y1 - 2004

N2 - This study develops a two-step estimator for a panel data Tobit model based on taking first-differences of the equation of interest, under conditional mean independence assumptions.The necessary correction terms are non-standard and a substantial part is therefore devoted to the formal derivation of these correction terms.The main advantage of this estimator is that it yields estimates that are far less sensitivity to misspecification of the conditional mean independence assumption than an estimation procedure set up in levels.Monte Carlo simulations are provided in support of this.

AB - This study develops a two-step estimator for a panel data Tobit model based on taking first-differences of the equation of interest, under conditional mean independence assumptions.The necessary correction terms are non-standard and a substantial part is therefore devoted to the formal derivation of these correction terms.The main advantage of this estimator is that it yields estimates that are far less sensitivity to misspecification of the conditional mean independence assumption than an estimation procedure set up in levels.Monte Carlo simulations are provided in support of this.

KW - panel data

KW - estimating

KW - monte carlo technique

M3 - Discussion paper

VL - 2004-67

T3 - CentER Discussion Paper

BT - A Two-Step First Difference Estimator for a Panel Data Tobit Model under Conditional Mean Independence Assumptions

PB - Econometrics

CY - Tilburg

ER -