Faithin Omaha eBook Collection

Download or read book Essays on Weak Identification, Model Selection and Hypothesis Testing in Econometrics written by Purevdorj Tuvaandorj and published by . This book was released on 2015 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: "This thesis makes contributions to weak identification, modelselection and hypothesis testing in econometrics. It consists of thefollowing essays.In Chapter 1, we study likelihood-basedinference in models with possible identification failure. The results relyheavily on the properties of the mapping from structural parameters togeneralized reduced-form parameters (which are identified by construction).We establish an asymptotic chi-square bound on the likelihood ratio (LR)statistic for testing restrictions on the possibly unidentified structuralparameters with degrees of freedom equal to the dimension of the reducedform parameter vector through which the tested parameters enter thelikelihood function. We also propose pivotal C(alpha)-type statisticsthat are robust to potential identification failure and are flexible inincorporating a wide class of estimators of the (strongly identified)nuisance parameters. Furthermore, we develop a generalized version of theclassical Anderson-Rubin (AR)-type statistic in linear simultaneousequations and an identification-robust pretest-based inference procedure.In Chapter 2, we study the invariance properties of various test criteria which have been proposed for hypothesis testing in the context of incompletely specified models, such asmodels which are formulated in terms of estimating functions (Godambe, 1960, Ann. Math. Stat.) or moment conditions and are estimated bygeneralized method of moments (GMM) procedures (Hansen, 1982, Econometrica), and models estimated by pseudo-likelihood (Gourieroux,Monfort and Trognon, 1984, Econometrica) and M-estimation methods.The invariance properties considered include invariance to (possiblynonlinear) hypothesis reformulations and reparameterizations. The teststatistics examined include Wald-type, LR-type, LM-type, score-type, and C(alpha)-type criteria. In Chapter 3, we propose generalized C(alpha) tests for testing linear and nonlinear parameterrestrictions in models specified by estimating functions. The asymptotic distribution of theproposed statistic is established under weak regularity conditions. We show that earlierC(alpha)-type statistics are included as special cases. The problem of testing hypotheses fixinga subvector of the complete parameter vector of the model is discussed in detail. In Chapter 4, we consider conditional distribution and conditional density functionalsin the space of generalized functions. We obtain the limit of the kernel estimators for weakly dependent data, evenunder non-differentiability of the distribution function; the limit Gaussian process is characterizedas a stochastic random functional (random generalized function) on the suitablefunction space. An alternative simple to compute estimator based on the empirical distribution function is proposed for the generalized random functional. For test statistics based on this estimator, limit properties are established.Chapter 5, considers the issue of selecting the number of regressors and the numberof structural breaks in multivariate regression models in the possible presence of multiplestructural changes. We develop a modified Akaike's information criterion (AIC), amodified Mallows' Cp criterion and a modified Bayesian information criterion (BIC). Thepenalty terms in these criteria are shown to be different from the usual terms." --