Estimation and Inference in Econometrics

by Davidson, Russell; Mackinnon, James G.

• ISBN13: 9780195060119
• ISBN10: 0195060113
• Format: Hardcover
• Copyright: 1993-01-14
• Publisher: Oxford University Press

Summary

Offering students a unifying theoretical perspective, this innovative textemphasizes nonlinear techniques of estimation, including nonlinear leastsquares, nonlinear instrumental variables, maximum likelihood and thegeneralized method of moments, but nevertheless relies heavily on simplegeometrical arguments to develop intuition. One theme of the book is the use ofartificial regressions for estimation, inference, and specification testing ofnonlinear models, including diagnostic tests for parameter constancy, seriescorrelation, heteroskedasticity and other types of misspecification. Othertopics include the linear simultaneous equations model, non-nested hypothesistests, influential observations and leverage, transformations of the dependentvariable, binary response models, models for time-series/cross-section data,multivariate models, seasonality, unit roots and cointegration, and Monte Carlomethods, always with an emphasis on problems that arise in applied work.Explaining throughout how estimates can be obtained and tests can be carriedout, the text goes beyond a mere algebraic description to one that can be easilytranslated into the commands of a standard econometric software package. Acomprehensive and coherent guide to the most vital topics in econometrics today,this text is indispensable for all levels of students of econometrics,economics, and statistics on regression and related topics.

Table of Contents

The Geometry of Least Squares

Nonlinear Regression Models and Nonlinear Least Squares

Inference in Nonlinear Regression Models

Introduction to Asymptotic Theory and Methods

Asymptotic Methods and Nonlinear Least Squares

The Gauss-Newton Regression

Instrumental Variables

The Method of Maximum Likelihood

Maximum Likelihood and Generalized Least Squares

Serial Correlation

Tests Based on the Gauss-Newton Regression

Interpreting Tests in Regression Directions

The Classical Hypothesis Tests

Transforming the Dependent Variable

Qualitative and Limited Dependent Variables

Heteroskedasticity and Related Topics

The Generalized Method of Moments

Simultaneous Equations Models

Regression Models for Time-series Data

Unit Roots and Cointegratiaon

Monte Carlo Experiments

Matrix Algebra

Results from Probability Theory

References

Author Index

Subject Index

Table of Contents provided by Publisher. All Rights Reserved.