Instrumental variable estimation with heteroskedasticity and many instruments

Jerry A. Hausman, Whitney K. Newey, Tiemen M Woutersen, John C. Chao, Norman R. Swanson

Research output: Contribution to journalArticle

35 Citations (Scopus)

Abstract

This paper gives a relatively simple, well behaved solution to the problem of many instruments in heteroskedastic data. Such settings are common in microeconometric applications where many instruments are used to improve efficiency and allowance for heteroskedasticity is generally important. The solution is a Fuller (1977) like estimator and standard errors that are robust to heteroskedasticity and many instruments. We show that the estimator has finite moments and high asymptotic efficiency in a range of cases. The standard errors are easy to compute, being like White's (1982), with additional terms that account for many instruments. They are consistent under standard, many instrument, and many weak instrument asymptotics. We find that the estimator is asymptotically as efficient as the limited-information maximum likelihood (LIML) estimator under many weak instruments. In Monte Carlo experiments, we find that the estimator performs as well as LIML or Fuller (1977) under homoskedasticity, and has much lower bias and dispersion under heteroskedasticity, in nearly all cases considered.

Original languageEnglish (US)
Pages (from-to)211-255
Number of pages45
JournalQuantitative Economics
Volume3
Issue number2
DOIs
StatePublished - Jul 2012

Fingerprint

Heteroskedasticity
Instrumental variable estimation
Estimator
Standard error
Weak instruments
Limited information
Maximum likelihood
Maximum likelihood estimator
Monte Carlo experiment
Microeconometrics
Asymptotic efficiency

Keywords

  • C12
  • C13
  • C23
  • Heteroskedasticity
  • Instrumental variables
  • Jackknife
  • Many instruments

ASJC Scopus subject areas

  • Economics and Econometrics

Cite this

Instrumental variable estimation with heteroskedasticity and many instruments. / Hausman, Jerry A.; Newey, Whitney K.; Woutersen, Tiemen M; Chao, John C.; Swanson, Norman R.

In: Quantitative Economics, Vol. 3, No. 2, 07.2012, p. 211-255.

Research output: Contribution to journalArticle

Hausman, Jerry A. ; Newey, Whitney K. ; Woutersen, Tiemen M ; Chao, John C. ; Swanson, Norman R. / Instrumental variable estimation with heteroskedasticity and many instruments. In: Quantitative Economics. 2012 ; Vol. 3, No. 2. pp. 211-255.
@article{e48b7753d8df4733b22a23abbd922f9b,
title = "Instrumental variable estimation with heteroskedasticity and many instruments",
abstract = "This paper gives a relatively simple, well behaved solution to the problem of many instruments in heteroskedastic data. Such settings are common in microeconometric applications where many instruments are used to improve efficiency and allowance for heteroskedasticity is generally important. The solution is a Fuller (1977) like estimator and standard errors that are robust to heteroskedasticity and many instruments. We show that the estimator has finite moments and high asymptotic efficiency in a range of cases. The standard errors are easy to compute, being like White's (1982), with additional terms that account for many instruments. They are consistent under standard, many instrument, and many weak instrument asymptotics. We find that the estimator is asymptotically as efficient as the limited-information maximum likelihood (LIML) estimator under many weak instruments. In Monte Carlo experiments, we find that the estimator performs as well as LIML or Fuller (1977) under homoskedasticity, and has much lower bias and dispersion under heteroskedasticity, in nearly all cases considered.",
keywords = "C12, C13, C23, Heteroskedasticity, Instrumental variables, Jackknife, Many instruments",
author = "Hausman, {Jerry A.} and Newey, {Whitney K.} and Woutersen, {Tiemen M} and Chao, {John C.} and Swanson, {Norman R.}",
year = "2012",
month = "7",
doi = "10.3982/QE89",
language = "English (US)",
volume = "3",
pages = "211--255",
journal = "Quantitative Economics",
issn = "1759-7323",
publisher = "The Economic Society",
number = "2",

}

TY - JOUR

T1 - Instrumental variable estimation with heteroskedasticity and many instruments

AU - Hausman, Jerry A.

AU - Newey, Whitney K.

AU - Woutersen, Tiemen M

AU - Chao, John C.

AU - Swanson, Norman R.

PY - 2012/7

Y1 - 2012/7

N2 - This paper gives a relatively simple, well behaved solution to the problem of many instruments in heteroskedastic data. Such settings are common in microeconometric applications where many instruments are used to improve efficiency and allowance for heteroskedasticity is generally important. The solution is a Fuller (1977) like estimator and standard errors that are robust to heteroskedasticity and many instruments. We show that the estimator has finite moments and high asymptotic efficiency in a range of cases. The standard errors are easy to compute, being like White's (1982), with additional terms that account for many instruments. They are consistent under standard, many instrument, and many weak instrument asymptotics. We find that the estimator is asymptotically as efficient as the limited-information maximum likelihood (LIML) estimator under many weak instruments. In Monte Carlo experiments, we find that the estimator performs as well as LIML or Fuller (1977) under homoskedasticity, and has much lower bias and dispersion under heteroskedasticity, in nearly all cases considered.

AB - This paper gives a relatively simple, well behaved solution to the problem of many instruments in heteroskedastic data. Such settings are common in microeconometric applications where many instruments are used to improve efficiency and allowance for heteroskedasticity is generally important. The solution is a Fuller (1977) like estimator and standard errors that are robust to heteroskedasticity and many instruments. We show that the estimator has finite moments and high asymptotic efficiency in a range of cases. The standard errors are easy to compute, being like White's (1982), with additional terms that account for many instruments. They are consistent under standard, many instrument, and many weak instrument asymptotics. We find that the estimator is asymptotically as efficient as the limited-information maximum likelihood (LIML) estimator under many weak instruments. In Monte Carlo experiments, we find that the estimator performs as well as LIML or Fuller (1977) under homoskedasticity, and has much lower bias and dispersion under heteroskedasticity, in nearly all cases considered.

KW - C12

KW - C13

KW - C23

KW - Heteroskedasticity

KW - Instrumental variables

KW - Jackknife

KW - Many instruments

UR - http://www.scopus.com/inward/record.url?scp=84873715704&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873715704&partnerID=8YFLogxK

U2 - 10.3982/QE89

DO - 10.3982/QE89

M3 - Article

VL - 3

SP - 211

EP - 255

JO - Quantitative Economics

JF - Quantitative Economics

SN - 1759-7323

IS - 2

ER -