A Robust Test for Weak Instruments in Stata

Pflueger, Carolin E.; Wang, Su

doi:10.2139/ssrn.2323012

A Robust Test for Weak Instruments in Stata

16 Pages Posted: 5 Nov 2013 Last revised: 3 Dec 2014

See all articles by Carolin E. Pflueger

Carolin E. Pflueger

University of Chicago - Harris Public Policy; National Bureau of Economic Research (NBER)

Su Wang

London School of Economics & Political Science (LSE)

Date Written: July 1, 2014

Abstract

We introduce and describe a Stata routine weakivtest implementing the test for weak instruments of Montiel Olea and Pflueger (2013). weakivtest allows for errors that are not conditionally homoskedastic and serially uncorrelated. It extends the Stock and Yogo (2005) weak instrument tests available in ivreg2 and in the ivregress postestimation command estat firststage.weakivtest tests the null hypothesis that instruments are weak or that the estimator Nagar bias is large relative to a benchmark for both Two-Stage Least Squares (TSLS) and Limited Information Maximum Likelihood (LIML) with one single endogenous regressor. The routine can accommodate Eicker-Huber-White heteroskedasticity robust, Newey and West (1987) heteroskedasticity- and autocorrelation-consistent, and clustered variance estimates.

Keywords: F Statistic, Heteroskedasticity, Autocorrelation, Clustered, Stata

JEL Classification: B23

Suggested Citation: Suggested Citation

Pflueger, Carolin E. and Wang, Su, A Robust Test for Weak Instruments in Stata (July 1, 2014). Available at SSRN: https://ssrn.com/abstract=2323012 or http://dx.doi.org/10.2139/ssrn.2323012