Heterogeneous Treatment Effects

Instrumental Variables Without Monotonicity?

Research output: Working paperDiscussion paperOther research output

233 Downloads (Pure)

Abstract

A fundamental identification problem in program evaluation arises when idiosyncratic gains from participation and the treatment decision depend on each other. Imbens and Angrist (1994) were the first to exploit a monotonicity condition in order to identify a local average treatment effect parameter using instrumental variables. More recently, Heckman and Vytlacil (1999) suggested estimation of a variety of treatment effect parameters using a local version of their approach. However, identification hinges on the same monotonicity assumption that is fundamentally untestable. We investigate the sensitivity of respective estimates to reasonable departures from monotonicity that are likely to be encountered in practice. Approximations to respective bias terms are derived. In an empirical application the bias is calculated and bias corrected estimates are obtained. The accuracy of the approximation is investigated in a Monte Carlo study.
Original languageEnglish
Place of PublicationTilburg
PublisherEconometrics
Number of pages39
Volume2008-45
Publication statusPublished - 2008

Publication series

NameCentER Discussion Paper
Volume2008-45

Fingerprint

Heterogeneous treatment effects
Monotonicity
Instrumental variables
Approximation
Heckman
Identification problem
Participation
Monte Carlo study
Treatment effects
Average treatment effect
Program evaluation

Keywords

  • Program evaluation
  • heterogeneity
  • identification
  • dummy endogenous variable
  • selection on unobservables
  • instrumental variables
  • monotonicity
  • nonseparable index selection model

Cite this

Klein, T. J. (2008). Heterogeneous Treatment Effects: Instrumental Variables Without Monotonicity? (CentER Discussion Paper; Vol. 2008-45). Tilburg: Econometrics.
Klein, T.J. / Heterogeneous Treatment Effects : Instrumental Variables Without Monotonicity?. Tilburg : Econometrics, 2008. (CentER Discussion Paper).
@techreport{e015611a6b974cd58dc87818c6c5e39e,
title = "Heterogeneous Treatment Effects: Instrumental Variables Without Monotonicity?",
abstract = "A fundamental identification problem in program evaluation arises when idiosyncratic gains from participation and the treatment decision depend on each other. Imbens and Angrist (1994) were the first to exploit a monotonicity condition in order to identify a local average treatment effect parameter using instrumental variables. More recently, Heckman and Vytlacil (1999) suggested estimation of a variety of treatment effect parameters using a local version of their approach. However, identification hinges on the same monotonicity assumption that is fundamentally untestable. We investigate the sensitivity of respective estimates to reasonable departures from monotonicity that are likely to be encountered in practice. Approximations to respective bias terms are derived. In an empirical application the bias is calculated and bias corrected estimates are obtained. The accuracy of the approximation is investigated in a Monte Carlo study.",
keywords = "Program evaluation, heterogeneity, identification, dummy endogenous variable, selection on unobservables, instrumental variables, monotonicity, nonseparable index selection model",
author = "T.J. Klein",
note = "Subsequently published in Journal of Econometrics, 2010 Pagination: 39",
year = "2008",
language = "English",
volume = "2008-45",
series = "CentER Discussion Paper",
publisher = "Econometrics",
type = "WorkingPaper",
institution = "Econometrics",

}

Klein, TJ 2008 'Heterogeneous Treatment Effects: Instrumental Variables Without Monotonicity?' CentER Discussion Paper, vol. 2008-45, Econometrics, Tilburg.

Heterogeneous Treatment Effects : Instrumental Variables Without Monotonicity? / Klein, T.J.

Tilburg : Econometrics, 2008. (CentER Discussion Paper; Vol. 2008-45).

Research output: Working paperDiscussion paperOther research output

TY - UNPB

T1 - Heterogeneous Treatment Effects

T2 - Instrumental Variables Without Monotonicity?

AU - Klein, T.J.

N1 - Subsequently published in Journal of Econometrics, 2010 Pagination: 39

PY - 2008

Y1 - 2008

N2 - A fundamental identification problem in program evaluation arises when idiosyncratic gains from participation and the treatment decision depend on each other. Imbens and Angrist (1994) were the first to exploit a monotonicity condition in order to identify a local average treatment effect parameter using instrumental variables. More recently, Heckman and Vytlacil (1999) suggested estimation of a variety of treatment effect parameters using a local version of their approach. However, identification hinges on the same monotonicity assumption that is fundamentally untestable. We investigate the sensitivity of respective estimates to reasonable departures from monotonicity that are likely to be encountered in practice. Approximations to respective bias terms are derived. In an empirical application the bias is calculated and bias corrected estimates are obtained. The accuracy of the approximation is investigated in a Monte Carlo study.

AB - A fundamental identification problem in program evaluation arises when idiosyncratic gains from participation and the treatment decision depend on each other. Imbens and Angrist (1994) were the first to exploit a monotonicity condition in order to identify a local average treatment effect parameter using instrumental variables. More recently, Heckman and Vytlacil (1999) suggested estimation of a variety of treatment effect parameters using a local version of their approach. However, identification hinges on the same monotonicity assumption that is fundamentally untestable. We investigate the sensitivity of respective estimates to reasonable departures from monotonicity that are likely to be encountered in practice. Approximations to respective bias terms are derived. In an empirical application the bias is calculated and bias corrected estimates are obtained. The accuracy of the approximation is investigated in a Monte Carlo study.

KW - Program evaluation

KW - heterogeneity

KW - identification

KW - dummy endogenous variable

KW - selection on unobservables

KW - instrumental variables

KW - monotonicity

KW - nonseparable index selection model

M3 - Discussion paper

VL - 2008-45

T3 - CentER Discussion Paper

BT - Heterogeneous Treatment Effects

PB - Econometrics

CY - Tilburg

ER -

Klein TJ. Heterogeneous Treatment Effects: Instrumental Variables Without Monotonicity? Tilburg: Econometrics. 2008. (CentER Discussion Paper).