A New Specification Test For The Validity Of Instrumental Variables

By Jinyong Hahn, Jerry A. Hausman. 2000

447 citations recorded by the source corpus.

We develop a new specification test for IV estimators adopting a particular second order approximation of Bekker. The new specification test compares the difference of the forward (conventional) 2SLS estimator of the coefficient of the right-hand side endogenous variable with the reverse 2SLS estimator of the same unknown parameter when the normalization is changed. Under the null hypothesis that conventional first order asymptotics provide a reliable guide to inference, the two estimates should be very similar. Our test sees whether the resulting difference in the two estimates satisfies the results of second order asymptotic theory. Essentially the same idea is applied to develop another new specification test using second-order unbiased estimators of the type first proposed by Nagar. If the forward and reverse Nagar-type estimators are not significantly different we recommend estimation by LIML, which we demonstrate is the optimal linear combination of the Nagar-type estimators (to second order). We also demonstrate the high degree of similarity for k-class estimators between the approach of Bekker and the Edgeworth expansion approach of Rothenberg. An empirical example and Monte Carlo evidence demonstrate the operation of the new specification test.

Subfields: Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/f4c734c1d17a7ff1cd019e5ebea66f614eb2d941

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Endogeneity and the corporate governance - performance relation (2010, background): “CAPEX (Capital Expenditure to Sales): CAPEX is included as a control for performance as it is expected that investment in capital will be positively related to the discounted value of future growth opportunities of the firm (Lang et al., 1989); 3. LEVERAGE: The capital structure of a firm is expecte”
Instrumental variable estimation in the presence of many moment conditions (2011, background): “Hahn and Hausman (2002) observe that the theoretical R(2) of the first stage regression is given by R(2) f = π ′π/(π′π+1).”
Controlling for endogeneity with instrumental variables in strategic management research (2008, methodology): “15 Other tests, such as the Hahn and Hausman (2002), Shea (1997) or canonical correlations test (Anderson, 1984; Hall et al., 1996), can lead to fallacious conclusions.”
The Impact of Free Movement of Workers from Central and Eastern Europe on the UK Labour Market (2006, methodology): “Nonetheless the Hausman test (Hahn and Hausman, 2002) suggest that there was no endogeneity in the model in the first place and the Hansen-Sargan test does not fail the instruments.”
Instrumental variable estimation based on grouped data (2005, methodology): “In particular, the specification tests of Hahn and Hausman (1999) can be expected to be valid only in the Gaussian case.”
Valuating the effects of beach erosion to tourism revenue. A management perspective (2015, methodology): “The validity of the instruments used as endogenous variables was tested at the first stage regression, through the squared partial correlation (R2) between the instruments (closure depth, presence of coastal road and sector length) and the endogenous regressor (beach width) (Hahn and Hausman, 2002; ”
Estimation With Many Instrumental Variables (2006, methodology): “Hahn and Hausman (2002) gave a test for weak instruments and Hahn, Hausman, and Kuersteiner (2004) showed that FULL performs well under weak instruments.”
Identification With Imperfect Instruments (2008, background): “Our results are consistent with the nding of Hahn and Hausman (2003a) that if E [ZU ] 6= 0, it is possible for OLS to be closer to than IV z .”
Finite sample evidence of IV estimators under weak instruments (2007, background): “The literature on the finite sample problems of IV estimators under weak instruments has grown after Nelson and Startz (1990a, 1990b) and Maddala and Jeong (1992) pointed out that, under weak instruments, two-stage least squares (TSLS) is severely biased in finite samples towards the expectation of ”
Dynamic Models for Short Panels (2008, background): “Moreover, there is a signi…cant amount of recent literature highlighting the deleterious e¤ects of weak instruments in general (see, for example Wang and Zivot, 1998, Woglom, 2001 and Hahn and Hausman, 2002 ).”