G. S. MADDALA Introduction to : Econometrics SECOND EDITIONG.S. MADDALA UNIVERSITY OF FLORIDA AND OHIO STATE UNIVERSITY Introduction to Econometrics SECOND EDITION MACMILLAN PUBLISHING COMPANY NEW YORK Maxwell Macmillan Canada TORONTO Maxwell Macmillan International NEW YORK OXFORD SINGAPORE SYDNEYCopyright © 1992 by Macmillan Publishing Company, a division of Macmillan, Inc. PRINTED IN THE UNITED STATES OF AMERICA All rights reserved. No part of this book may be reproduced or transmitted in any form ot by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Earlier edition copyright © 1988 by Macmillan Publishing Company Macmillan Publishing Company 866 Third Avenue, New York, New York 10022 Macmillan Publishing Company is part of the Maxwell Communication Group of Companies. Maxwell Macmillan Canada, Inc. 1200 Eglinton Avenue East Suite 200 Don Mills, Ontario M3C 3N1 LIBRARY OF CONGRESS CATALOGING IN PUBLICATION DATA Maddala, G. S. Introduction to econometrics / G. S. Maddala. — 2nd ed. p. cm, Includes index ISBN 0-02-374545-2 1. Econometrics. I. Title. HB139.M353_ 1992 330°.01°S195—de20 91-23860 clip Printing: 2345678 Year: 23456789Contents Preface xv 1 What Is Econometrics? 1 2 1.1. What Is Econometrics? = 1 1.2. Economic and Econometric Models 2 1.3. The Aims and Methodology of Econometrics 4 1.4 What Constitutes a Test of an Economic Theory? 6 Summary and an Outline of the Book 8 Statistical Background and Matrix Algebra 11 2.1 Introduction 12 2.2. Probability 12 2.3 Random Variables and Probability Distributions — 17 2.4 The Normal Probability Distribution and Related Distributions 19 2.5 Classical Statistical Inference 21 2.6 Properties of Estimators 23 2.7. Sampling Distributions for Samples from a Normal Population 26viii CONTENTS 2.8 Interval Estimation 27 2.9 Testing of Hypotheses 28 2.10 Relationship Between Confidence Interval Procedures and Tests of Hypotheses 32 Summary 33 Exercises 34 Appendix to Chapter2 41 3 Simple Regression 59 3.1 Introduction 59 3.2 Specification of the Relationships 62. 3.3. The Method of Moments 66 3.4 Method of Least Squares 68 3.5 Statistical Inference in the Linear Regression Model 76 3.6 Analysis of Variance for the Simple Regression Model 84 3.7 Prediction with the Simple Regression Model 85 3.8 Outliers 88 3.9 Alternative Functional Forms for Regression Equations 96 *3.10 Inverse Prediction in the Least Squares Regression Model 101 *3.11 Stochastic Regressors 103 *3,12 The Regression Fallacy 104 Summary 106 Exercises 108 Appendix to Chapter3 114 4 Multiple Regression 127 Introduction 128 A Model with Two Explanatory Variables 129 Statistical Inference in the Multiple Regression Model 134 4.4. Interpretation of the Regression Coefficients 143 4.5 Partial Correlations and Multiple Correlation 146 PPP BRECONTENTS. 46 47 48 49 Relationships Among Simple, Partial, and Multiple Correlation Coefficients 147 Prediction in the Multiple Regression Model — 154 Analysis of Variance and Tests of Hypotheses 156 Omission of Relevant Variables and Inclusion of Irrelevant Variables 161 4.10 Degrees of Freedom and R? 165 4.11 Tests for Stability 170 “4.12 The LR, W, and LM Tests 177 Summary 179 Exercises 181 Appendix to Chapter 4 187 Data Sets 194 5 Heteroskedasticity 201 Introduction 201 Detection of Heteroskedasticity 203 Consequences of Heteroskedasticity 209 Solutions to the Heteroskedasticity Problem — 212 Heteroskedasticity and the Use of Deflators 215 Testing the Linear Versus Log-Linear Functional Form 220 Summary 223 Exercises 224 Appendix to Chapter 5 227 6 Autocorrelation 229 6.1 Introduction 229 6.2, Durbin-Watson Test 230 6.3 Estimation in Levels Versus First Differences 232 6.4 Estimation Procedures with Autocorrelated Errors 236 6.5 Effect of AR(I) Errors on OLS Estimates 241 6.6 Some Further Comments on the DW Test 245 6.7 Tests for Serial Correlation in Models with Lagged Dependent Variables 248x CONTENTS 6.8 A General Test for Higher-Order Serial Correlation: The LM Test = 250 6.9 Strategies When the DW Test Statistic Is Significant 252 %6.10 Trends and Random Walks = 258, *6.11 ARCH Models and Serial Correlation 264 Summary 265 Exercises 267 7 Multicollinearity 269 7.1 Introduction 269 7.2 Some Illustrative Examples 270 7.3, Some Measures of Multicollinearity 274 7.4 Problems with Measuring Multicollinearity 276 7.5 Solutions to the Multicollinearity Problem: Ridge Regression 280 7.6 Principal Component Regression 284 7.7 Dropping Variables 289 7.8 Miscellaneous Other Solutions — 292 Summary 294 Exercises 295 Appendix to Chapter 7 296 8 Dummy Variables and Truncated Variables 305 8.1 Introduction 306 8.2. Dummy Variables for Changes in the Intercept Term 306 8.3 Dummy Variables for Changes in Slope Coefficients 312 8.4 Dummy Variables for Cross-Equation Constraints 315 8.5 Dummy Variables for Testing Stability of Regression Coefficients 318 8.6 Dummy Variables Under Heteroskedasticity and Autocorrelation — 320 8.7 Dummy Dependent Variables 322CONTENTS: 9 xi 8.8 The Linear Probability Model and the Linear Discriminant Function 323 8.9 The Probit and Logit Models 327 8.10 Illustrative Example 335 8.11 Truncated Variables: The Tobit Model 338 Summary 344 Exercises 345 Simultaneous Equations Models 355 10 9.1 Introduction 356 9.2 Endogenous and Exogenous Variables 357 9.3 The Identification Problem: Identification Through Reduced Form 358 9.4 Necessary and Sufficient Conditions for Identification 363 9.5 Methods of Estimation: The Instrumental Variable Method 366 9.6 Methods of Estimation: The Two-Stage Least Squares Method 373 9.7 The Question of Normalization 377 *9.8 The Limited-Information Maximum Likelihood Method 381 *9.9 On the Use of OLS in the Estimation of Simultaneous-Equations Models 383 ¥9.10 Exogeneity and Causality 389 Summary 395 Exercises 397 Appendix to Chapter9 400 Models of Expectations 405 10.1 Introduction 405 10.2. Naive Models of Expectations 406 10.3 The Adaptive Expectations Model 408 10.4 Estimation with the Adaptive Expectations Model 410xii CONTENTS 10.5 Two Illustrative Examples 413 10.6 Expectational Variables and Adjustment Lags 415 10.7 Partial Adjustment with Adaptive Expectations 421 10.8 Alternative Distributed Lag Models: Polynomial Lags 423 10.9 Rational Lags 429 10.10 Rational Expectations 431 10.11 Tests for Rationality 434 10.12 Estimation of a Demand and Supply Model Under Rational Expectations 436 10.13 The Serial Correlation Problem in Rational Expectations Models 443 Summary 444 Exercises 445 11 Errors in Variables 447 11.1 Introduction 447 The Classical Solution for a Single-Equation Model with One Explanatory Variable 448 11.3 The Single-Equation Model with Two Explanatory Variables 451 11.4 Reverse Regression 459 11.5 Instrumental Variable Methods 461 11.6 Proxy Variables 464 11.7 Some Other Problems 468 Summary 470 Exercises 472 12 Diagnostic Checking, Model Selection, and Specification Testing 475 12.1 Introduction 476 Diagnostic Tests Based on Least Squares Residuals 477 12.3. Problems with Least Squares Residuals 479 12.4 Some Other Types of Residuals 481CONTENTS xiii 12.5 DFFITS and Bounded Influence Estimation 487 12.6 Model Selection 490 12.7 Selection of Regressors 496 12.8 Implied F-Ratios for the Various Criteria 500 12.9 Cross-Validation 504 12.10 Hausman’s Specification Error Test 506 12,11 The Plosser-Schwert-White Differencing Test 12.12 Tests for Nonnested Hypotheses 514 Summary 518 Exercises 521 Appendix to Chapter 12 523 13 513 Introduction to Time-Series Analysis 13.1 Introduction 525 13.2. Two Methods of Time-Series Analysis: Frequenc’ Domain and Time Domain 526 13.3 Stationary and Nonstationary Time Series 527 13.4 Some Useful Models for Time Series 530 525 y 13.5 Estimation of AR, MA, and ARMA Models 537 13.6 The Box-Jenkins Approach 542 13.7 R? Measures in Time-Series Models 549 Summary 553 Exercises 553 Data Sets 555 14 Vector Autoregressions, Unit Roots, and Cointegration 14.1 Introduction $78 14.2 Vector Autoregressions 578 14,3 Problems with VAR Models in Practice 580 14.4 Unit Roots 581 14,5 Unit Root Tests 582 14.6 Cointegration 588 14.7 The Cointegrating Regression 590 14.8 Vector Autoregressions and Cointegration 592 577