Econometrics (Q5906679)

This volume 11 of the ``Handbook of Statistics'' is concerned with advanced econometric techniques. The editors divided the Handbook into seven distinct parts with three to six chapters for each part. The overall impression is that each chapter is written by experts in this field and that the different contributions are of a high scientific standard, covering most of the latest developments in these areas. Many aspects or special developments in certain research areas, published in several journals only up to now, are tied together and provide a comprehensive, unique overview to the interested reader. The first part deals with ``Endogeneous stratifications, semiparametric and nonparametric estimation'' and covers topics such as the estimation from endogeneously stratified samples, including logistic regressions and semiparametric models of discrete choice. With respect to quantal response models, \textit{J. L. Horowitz} discussed alternative semiparametric as well as nonparametric estimation. Properties of the different estimators are given and an application demonstrates the usefulness of these approaches. Of a more general nature are the contributions by \textit{C. F. Manski} (The selection problem in econometrics and statistics) and by \textit{A. Ullah} and \textit{H. D. Vinod} (General nonparametric regression estimation and testing in econometrics). Part II is dedicated to the case of limited-dependent variables in econometrics and deals with three different aspects: The first one by \textit{R. W. Blundell} and \textit{R. J. Smith} summarizes recent developments in the field of simultaneous microeconometric models with censored or qualitative dependent variables, the middle part by \textit{L.-F. Lee} generalizes the Tobit-model to the multivariate case and shows how to estimate such models as well as how to perform certain specification tests. Finally, \textit{G. S. Maddala} discusses estimation methods of limited dependent variable models under the special assumption that rational expectations prevail. The Time-Series Analysis part surveys nonlinear time series in macroeconomics with special emphasis on chaos and stochastic nonlinearities, structural time series models and the analysis of time series which are subject to changes in regime. Parts IV and V deal with likelihood methods and Bayesian inference on the one hand and with alternatives to likelihood methods on the other. With respect to the former, recent developments in Rao's score test are presented by \textit{R. Mukerjee} and pseudo-likelihood methods are introduced by \textit{C. Gourieroux} and \textit{A. Monfort}. As far as the alternatives are concerned, the Generalized Method of Moments absorbs most of the space (4 chapters). Further alternatives are the robust \(M\)- estimator, for which strong consistency under a general discrepancy function is considered, and as an aside the chapter ``Testing for heteroskedasticity'', including topics such as testing for ARCH-effects and estimation of ARCH in mean and \(G\)-Arch-models. Computer-intensive methods (part vi) cover the fields of simulation and bootstrapping. Aspects of simulation estimation methods are related to the case of limited-dependent variables and to the more special case of panel data models with limited dependent variables. Stochastic simulations for inference in nonlinear errors-in-variables models are discussed by \textit{J. Jeong} and \textit{G. S. Maddala}. Bootstrap methods go along with applications in econometrics. The final part contains several contributions which are somewhat mixed. These are the identification of outliers, statistical aspects of calibration in macroeconomics, issues concerned with panel-data models with rational expectations and, last not least, continuous time financial models. Summing up, the topics presented in this Handbook are up to date and allow the reader to refresh his present knowledge of the state of econometrics in a very understandable way.