Misspecification tests play an important role in detecting unreliable and inadequate economic models. This book brings together many results from the growing literature in econometrics on misspecification testing. It provides theoretical analyses and convenient methods for application. The main emphasis is on the Lagrange multiplier principle, which provides considerable unification, although several other approaches are also considered. The author also examines general checks for model adequacy that do not involve formulation of an alternative hypothesis. General and specific tests are discussed in the context of multiple regression models, systems of simultaneous equations, and models with qualitative or limited dependent variables.
• Brings together results from diverse literature on the tests for unreliable and inadequate economic models • Godfrey is one of the leading scholars in the field
Preface; 1. Approaches to testing the hypothesis of adequate specification; 2. Inequalities between criteria for testing hypotheses in linear regression models; 3. The Lagrange multiplier test and testing for misspecification: an extended analysis; 4. Tests for misspecification of regression equations; 5. Tests for misspecification of simultaneous equation models; 6. Tests for qualitative and limited dependent variable models; Bibliography; Index.
'An admirable attempt to provide a unified picture of the issues that have been raised in the testing literature, and to guide practitioners through the maze of different tests, to which they might wish to subject their models.' Russell Davidson, Journal of Economic Literature
'This book is a welcome addition to the Econometric Society Monograph Series and provides a comprehensive account of the Lagrange multiplier approach and other likelihood-based procedures to the problem of misspecification testing in econometrics.' M. H. Pesaren, The Economic Journal
'Written by one of the leading scholars in the field, the book is strongly recommended as an addition to the bookshelf especially for applied econometricians.' Joachim Wagner, Universität Hannover