A robust instrumental-variables estimator
Abstract. The classical instrumental-variables estimator is extremely sensitive
to the presence of outliers in the sample. This is a concern because outliers can
strongly distort the estimated effect of a given regressor on the dependent variable.
Although outlier diagnostics exist, they frequently fail to detect atypical
observations because they are themselves based on nonrobust (to outliers) estimators.
Furthermore, they do not take into account the combined influence
of outliers in the first and second stages of the instrumental-variables estimator.
In this article, we present a robust instrumental-variables estimator, initially
proposed by Cohen Freue, Ortiz-Molina, and Zamar (2011, Working paper:
http://www.stat.ubc.ca/˜ruben/website/cv/cohen-zamar.pdf ), that we have programmed
in Stata and made available via the robivreg command. We have improved
on their estimator in two different ways. First, we use a weighting scheme
that makes our estimator more efficient and allows the computations of the usual
identification and overidentifying restrictions tests. Second, we implement a generalized
Hausman test for the presence of outliers.
University of Strathclyde
University of Namur
(Centre for Research in the Economics of Development)
and Université Libre de Bruxelles
(European Center for Advanced Research in Economics and Statistics
and Center for Knowledge Economics)
View all articles by these authors:
Rodolphe Desbordes, Vincenzo Verardi
View all articles with these keywords:
robivreg, multivariate outliers, robustness, S-estimator, instrumental variables
Download citation: BibTeX RIS
Download citation and abstract: BibTeX RIS