Estimation with Weak Instruments: Accuracy of Higher-Order Bias and MSE Approximations

By Jinyong Hahn Brown, G. Kuersteiner. 2004

240 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/636af2b8a6e3b8934536edd54467a748989efb3a

Citations in the corpus

• A review of instrumental variable estimators for Mendelian randomization (2015, unknown): “Use of LIML has been strongly discouraged by some, as LIML estimates do not have defined moments for any number of instruments.(54) However, use has also been encouraged by others, especially with weak instruments, as the median of the distribution of the LIML estimator is close to unbiased even wit”
• Improved JIVE Estimators for Overidentified Linear Models with and without Heteroskedasticity (2009, unknown): “First note that with homoskedasticity, (16) is zero for both C I J I V E and CN AGAR .”
• Educational mismatch, wages, and wage growth: Overeducation in Sweden, 1974-2000 (2009, unknown): “However, it has been suggested that the bias in the point estimates may be dealt with by estimating the model using Fuller LIML (Andrews and Stock, 2005; Hahn et al., 2004; Hansen et al., 2005).”
• Massachusetts Institute of Technology Department of Economics Working Paper Series Sparse Models and Methods for Optimal Instruments with an Application to Eminent Domain Sparse Models and Methods for Optimal Instruments with an Application to Eminent Domain (unknown): “See [26] for additional discussion.”
• Servitization: Disentangling the Impact of Service Business Model Innovation on Manufacturing Firm Performance (2013, unknown): “In all models, we use one of the most common approaches to address endogeneity (Hahn et al., 2004; Murray, 2006) – the instrumental variables approach (Wooldridge, 2002).”
• Testing with many weak instruments (2007, unknown): “Many IV asymptotics (with strong IVs) have been employed in Anderson (1976), Kunitomo (1980), Morimune (1983), Bekker (1994), Donald and Newey (2001), Hahn (2002), Hahn et al. (2004), and Hansen et al. (2005) among others.”
• Controlling for endogeneity with instrumental variables in strategic management research (2008, unknown): “Two values, called , have been retained in practice: 1 and 4 (e.g. Davidson and McKinnon, 1993; Fuller, 1977; Hahn et al., 2004).”
• THE CASE AGAINST JIVE (2006, unknown): “where PW ≡ W (W>W )−1W> projects on to S(W ), the subspace spanned by the columns of W. The usual covariance matrix estimator is σ̂(2) IV(X >PW X)−1, where σ̂(2) IV ≡ 1 −n‖y −Xβ̂ IV‖2. (5) A classic alternative to the 2SLS estimator is the limited-information maximum likelihood, or LIML, estimator, ”
• Sparse Models and Methods for Optimal Instruments With an Application to Eminent Domain (2012, unknown): “See Hahn, Hausman, and Kuersteiner (2004) for additional discussion.”
• Schooling, educational achievement, and the Latin American growth puzzle (2012, unknown): “…the bottom part of Table 4 also reports results on the Fuller (1977) modification of the limited information maximum likelihood (LIML) estimator, which is more robust than 2SLS in the presence of weak instruments and performs relatively well in the simulation exercise of Hahn et al. (2004).”
• Asymptotic properties of the Hahn-Hausman test for weak-instruments (2005, unknown): “…estimator and its 3 An estimator that could be used in the event that the HH statistic rejects is OLS, which can have lower MSE than TSLS if the instruments are invalid (see the discussion in Hahn and Hausman (2003b)), however here we do not examine inference conditional on failing the HH pretest.”
• Does traffic congestion reduce employment growth (2009, unknown): “The instrumental variables regressions use LIML, because studies have shown it to perform better than two-stage least squares in the presence of weak instruments (Hahn et al., 2004; Stock and Yogo, 2005).”
• The Hausman test, and some alternatives, with heteroskedastic data (2012, unknown): “” Hall, Rudebusch and Wilcox (1996) recommend that we do not rely on asymptotic theory but focus on an analysis of the distributions of the estimated parameters and hypothesis tests conditional on the realized value of the relevance statistic. In our case we should take into account the sample corre”
• Enhanced routines for instrumental variables/GMM estimation and testing (2007, unknown): “In particular, there is evidence suggesting that CUE and LIML perform better than IV-GMM in the presence of weak instruments (Hahn et al. (2004)).”