Introduction to

Econometrics
Third Edition

G.S. Maddala
Formerly of Ohio State University

JOHN WILEY & SONS, LTD

Chichester • New York • Weinheim • Brisbane • Toronto • Singapore
 Contents

Foreword xvii
Preface to the Second Edition xix
Preface to the Third Edition xxiii
Obituary xxv

PART I INTRODUCTION AND THE LINEAR REGRESSION MODEL 1

1 What is Econometrics? 3
What is in this Chapter? 3
1.1 What is Econometrics? 3
1.2 Economic and Econometric Models 4
1.3 The Aims and Methodology of Econometrics 6
1.4 What Constitutes a Test of an Economic Theory? 9
Summary and an Outline of the Book 9

2 Statistical Background and Matrix Algebra 11

What is in this Chapter? 11
2.1 Introduction 12
2.2 Probability 12
Addition Rules of Probability 13
Conditional Probability and the Multiplication Rule 14
Bayes' Theorem 15
Summation and Product Operations 15
2.3 Random Variables and Probability Distributions 17
Joint, Marginal, and Conditional Distributions 18
Illustrative Example 18
2.4 The Normal Probability Distribution and Related Distributions 19
The Normal Distribution 19
Related Distributions 20
 CONTENTS

2.5 Classical Statistical Inference 21

Point Estimation 22
2.6 Properties of Estimators 23
Unbiasedness 23
Efficiency 24
Consistency 24
Other Asymptotic Properties 25
2.7 Sampling Distributions for Samples from a Normal Population 26
2.8 Interval Estimation 27
2.9 Testing of Hypotheses 28
2.10 Relationship Between Confidence Interval Procedures and Tests
of Hypotheses 32
2.11 Combining Independent Tests 33
Summary 33
Exercises 34
Appendix to Chapter 2 41
Matrix Algebra 41
Exercises on Matrix Algebra 56

Simple Regression 59
What is in this Chapter? 59
3.1 Introduction 59
3.2 Specification of the Relationships 61
3.3 The Method of Moments 65
Illustrative Example 66
3.4 The Method of Least Squares 68
Reverse Regression 71
Illustrative Example 72
3.5 Statistical Inference in the Linear Regression Model 75
Illustrative Example 77
Confidence Intervals for a, /S, and a1 78
Testing of Hypotheses 79
Example of Comparing Test Scores from the GRE and
GMAT Tests 81
Regression with No Constant Term 82
3.6 Analysis of Variance for the Simple Regression Model 83
3.7 Prediction with the Simple Regression Model 84
Prediction of Expected Values 86
Illustrative Example 87
3.8 Outliers 88
Some Illustrative Examples 89
3.9 Alternative Functional Forms for Regression Equations 94
Illustrative Example 97
*3.10 Inverse Prediction in the Least Squares Regression Model 99
*3.11 Stochastic Regressors 101
CONTENTS vii

*3.12 The Regression Fallacy 102

The Bivariate Normal Distribution 102
Galton's Result and the Regression Fallacy 104
A Note on the Term: "Regression" 104
Summary 105
Exercises 106
Appendix to Chapter 3 112

4 Multiple Regression 127

What is in this Chapter? 127
4.1 Introduction 127
4.2 A Model with Two Explanatory Variables 129
The Least Squares Method 130
Illustrative Example 132
4.3 Statistical Inference in the Multiple Regression Model 134
Illustrative Example 135
Formulas for the General Case of k Explanatory Variables 139
Some Illustrative Examples 140
4.4 Interpretation of the Regression Coefficients 143
Illustrative Example 145
4.5 Partial Correlations and Multiple Correlation 146
4.6 Relationships Among Simple, Partial, and Multiple Correlation
Coefficients 147
Two Illustrative Examples 148
4.7 Prediction in the Multiple Regression Model 153
Illustrative Example 153
4.8 Analysis of Variance and Tests of Hypotheses 154
Nested and Nonnested Hypotheses 156
Tests for Linear Functions of Parameters 157
Illustrative Example 158
4.9 Omission of Relevant Variables and Inclusion of Irrelevant
Variables 159
Omission of Relevant Variables 160
Example 1: Demand for Food in the United States 161
Example 2: Production Functions and Management Bias 162
Inclusion of Irrelevant Variables 163
4.10 Degrees of Freedom and R 164
4.11 Tests for Stability 168
The Analysis of Variance Test 168
Example 1: Stability of the Demand for Food Function 169
Example 2: Stability of Production Functions 170
Predictive Tests for Stability 173
Illustrative Example 173
*4.12 The LR, W, and LM Tests 176
Illustrative Example 176
viii CONTENTS

Summary 177
Exercises 179
Appendix to Chapter 4 185
The Multiple Regression Model in Matrix Notation 185
Data Sets 192

PART II VIOLATION OF THE ASSUMPTIONS OF THE BASIC MODEL 197

5 Heteroskedasticity 199
What is in this Chapter? 199
5.1 Introduction 199
Illustrative Example 200
5.2 Detection of Heteroskedasticity 202
Illustrative Example 202
Some Other Tests 203
Illustrative Example 205
An Intuitive Justification for the Breusch-Pagan Test 206
5.3 Consequences of Heteroskedasticity 207
Estimation of the Variance of the OLS Estimator Under
Heteroskedasticity 209
5.4 Solutions to the Heteroskedasticity Problem 209
Illustrative Example 211
5.5 Heteroskedasticity and the Use of Deflators 212
Illustrative Example: The Density Gradient Model 215
*5.6 Testing the Linear Versus Log-Linear Functional Form 217
The Box-Cox Test 217
The BM Test 219
The PE Test 219
Summary 220
Exercises 221
Appendix to Chapter 5 224
Generalized Least Squares 224

6 Autocorrelation 227
What is in this Chapter? 227
6.1 Introduction . 227
6.2 Durbin-Watson Test 228
Illustrative Example 229
6.3 Estimation in Levels Versus First Differences 230
Some Illustrative Examples 232
6.4 Estimation Procedures with Autocorrelated Errors 234
Iterative Procedures 236
Grid-Search Procedures 237
Illustrative Example 238
6.5 Effect of AR(1) Errors on OLS Estimates 238
CONTENTS ix

6.6 Some Further Comments on the DW Test 242

The von Neumann Ratio 243
The Berenblut-Webb Test 243
6.7 Tests for Serial Correlation in Models with Lagged
Dependent Variables 245
Durbin's h-Test 246
Durbin's Alternative Test 246
Illustrative Example 247
6.8 A General Test for Higher-Order Serial Correlation: The LM Test 248
6.9 Strategies When the DW Test Statistic is Significant 249
Errors Not AR(1) 249
Autocorrelation Caused by Omitted Variables 250
Serial Correlation Due to Misspecified Dynamics 252
The Wald Test 253
Illustrative Example 254
*6.10 Trends and Random Walks 255
Spurious Trends 257
Differencing and Long-Run Effects: The Concept of
Cointegration 258
*6.11 ARCH Models and Serial Correlation 260
6.12 Some Comments on the DW Test and Durbin's /i-Test and f-Test 262
Summary 262
Exercises 264

7 Multicollinearity 267
What is in this Chapter? 267
7.1 Introduction 268
7.2 Some Illustrative Examples 268
7.3 Some Measures of Multicollinearity 272
7.4 Problems with Measuring Multicollinearity 274
7.5 Solutions to the Multicollinearity Problem: Ridge Regression 278
7.6 Principal Component Regression 281
7.7 Dropping Variables 286
7.8 Miscellaneous Other Solutions 289
Using Ratios or First Differences 289
Using Extraneous Estimates 289
Getting More Data 291
Summary 291
Exercises 291
Appendix to Chapter 7 293
Linearly Dependent Explanatory Variables 293

8 Dummy Variables and Truncated Variables 301

What is in this Chapter? 301
8.1 Introduction 301
 CONTENTS

8.2 Dummy Variables for Changes in the Intercept Term 302

Illustrative Example 305
Two More Illustrative Examples 306
8.3 Dummy Variables for Changes in Slope Coefficients 307
8.4 Dummy Variables for Cross-Equation Constraints 310
8.5 Dummy Variables for Testing Stability of Regression
Coefficients 313
8.6 Dummy Variables Under Heteroskedasticity and
Autocorrelation 316
8.7 Dummy Dependent Variables 317
8.8 The Linear Probability Model and the Linear Discriminant
Function 318
The Linear Probability Model 318
The Linear Discriminant Function 320
8.9 The Probit and Logit Models 322
Illustrative Example 324
The Problem of Disproportionate Sampling 325
Prediction of Effects of Changes in the Explanatory Variables 327
Measuring Goodness of Fit 327
8.10 Illustrative Example 329
8.11 Truncated Variables: The Tobit Model 333
Some Examples 333
Method of Estimation 334
Limitations of the Tobit Model 335
The Truncated Regression Model 336
Summary 338
Exercises 339

Simultaneous Equations Models 343

What is in this Chapter? 343
9.1 Introduction 343
9.2 Endogenous and Exogenous Variables 345
9.3 The Identification Problem: Identification through Reduced Form 346
Illustrative Example 348
9.4 Necessary and Sufficient Conditions for Identification 351
Illustrative Example 353
9.5 Methods of Estimation: The Instrumental Variable Method 354
Measuring R2 356
Illustrative Example 3 357
9.6 Methods of Estimation: The Two-Stage Least Squares Method 360
Computing Standard Errors - 361
Illustrative Example 363
9.7 The Question of Normalization 366
*9.8 The Limited-Information Maximum Likelihood Method 367
Illustrative Example 368
CONTENTS xi

*9.9 On the Use of OLS in the Estimation of Simultaneous

Equations Models 369
Working's Concept of Identification 371
Recursive Systems 373
Estimation of Cobb-Douglas Production Functions 373
*9.10 Exogeneity and Causality 375
Weak Exogeneity 378
Superexogeneity 378
Strong Exogeneity 378
Granger Causality 379
Granger Causality and Exogeneity 380
Tests for Exogeneity 380
9.11 Some Problems with Instrumental Variable Methods 381
Summary 382
Exercises 383
Appendix to Chapter 9 386

10 Nonlinear Regressions, Models of Expectations, and Nonnormality 391

What is in this Chapter? 391
10.1 Introduction 392
10.2 The Newton-Raphson Method 392
10.3 Nonlinear Least Squares 393
The Gauss-Newton Method 393
10.4 Models of Expectations 394
10.5 Naive Models of Expectations 395
10.6 The Adaptive Expectations Model 397
10.7 Estimation with the Adaptive Expectations Model 399
Estimation in the Autoregressive Form 399
Estimation in the Distributed Lag Form 400
10.8 Two Illustrative Examples 401
10.9 Expectational Variables and Adjustment Lags 405
10.10 Partial Adjustment with Adaptive Expectations 409
10.11 Alternative Distributed Lag Models: Polynomial Lags 411
Finite Lags: The Polynomial Lag 412
Illustrative Example 415
Choosing the Degree of the Polynomial 416
10.12 Rational Lags 417
10.13 Rational Expectations 419
10.14 Tests for Rationality 422
10.15 Estimation of a Demand and Supply Model Under Rational
Expectations 424
Case 1 424
Case 2 425
Illustrative Example 428
10.16 The Serial Correlation Problem in Rational Expectations Models 431
xii CONTENTS

10.17 Nonnormality of Errors 431

Tests for Normality 432
10.18 Data Transformations 433
Summary 433
Exercises 435

11 Errors in Variables 437

What is in this Chapter? 437
11.1 Introduction 437
11.2 The Classical Solution for a Single-Equation Model with One
Explanatory Variable 438
11.3 The Single-Equation Model with Two Explanatory Variables 441
Two Explanatory Variables: One Measured with Error 441
Illustrative Example 444
Two Explanatory Variables: Both Measured with Error 446
11.4 Reverse Regression 449
11.5 Instrumental Variable Methods 451
11.6 Proxy Variables 454
Coefficient of the Proxy Variable 456
11.7 Some Other Problems 457
The Case of Multiple Equations 458
Correlated Errors 459
Summary 459
Exercises 461

PART III SPECIAL TOPICS 463

12 Diagnostic Checking, Model Selection, and Specification Testing 465

What is in this Chapter? 465
12.1 Introduction 465
12.2 Diagnostic Tests Based on Least Squares Residuals 466
Tests for Omitted Variables 467
Tests for ARCH Effects 468
12.3 Problems with Least Squares Residuals 469
12.4 Some Other Types of Residuals 470
Predicted Residuals, and Studentized Residuals 470
Dummy Variable Method for Studentized Residuals 471
BLUS Residuals 472
Recursive Residuals 472
Illustrative Example 474
12.5 DFFITS and Bounded Influence Estimation 476
Illustrative Example 478
12.6 Model Selection 479
Hypothesis-Testing Search 480
Interpretive Search 481
CONTENTS xiii

Simplification Search 481

Proxy Variable Search 481
Data Selection Search 482
Post-Data Model Construction 482
Hendry's Approach to Model Selection 483
12.7 Selection of Regressors 484
Theil's R 2 Criterion 486
Criteria Based on Minimizing the Mean-Squared
Error of Prediction 486
Akaike's Information Criterion 488
12.8 Implied F-Ratios for the Various Criteria 488
Bayes' Theorem and Posterior Odds for Model Selection 491
12.9 Cross-Validation 492
12.10 Hausman's Specification Error Test 494
An Application: Testing for Errors in Variables or Exogeneity 496
Some Illustrative Examples 497
An Omitted Variable Interpretation of the Hausman Test 498
12.11 The Plosser-Schwert-White Differencing Test 501
12.12 Tests for Nonnested Hypotheses 502
The Davidson and MacKinnon Test 502
The Encompassing Test 505
A Basic Problem in Testing Nonnested Hypotheses 506
Hypothesis Testing Versus Model Selection as a Research
Strategy 506
Summary 506
Exercises 508
Appendix to Chapter 12 510

13 Introduction to Time-Series Analysis 513

What is in this Chapter? 513
13.1 Introduction 513
13.2 Two Methods of Time-Series Analysis: Frequency Domain and
Time Domain 514
13.3 Stationary and Nonstationary Time Series 514
Strict Stationarity 515
Weak Stationarity 516
Properties of Autocorrelation Function 517
Nonstationarity 517
13.4 Some Useful Models for Time Series 517
Purely Random Process 517
Random Walk 518
Moving Average Process 519
Autoregressive Process 520
Autoregressive Moving Average Process 522
Autoregressive Integrated Moving Average Process 524
xiv CONTENTS

13.5 Estimation of AR, MA, and ARMA Models 524

Estimation of MA Models 524
Estimation of ARMA Models 525
Residuals from the ARMA Models 526
Testing Goodness of Fit 527
13.6 The Box-Jenkins Approach 529
Forecasting from Box-Jenkins Models 531
Illustrative Example 532
Trend Elimination: The Traditional Method 534
A Summary Assessment 535
Seasonality in the Box-Jenkins Modeling 535
13.7 R2 Measures in Time-Series Models 536
Summary 540
Exercises 540
Data Sets 541

14 Vector Autoregressions, Unit Roots, and Cointegration 543

What is in this Chapter? 543
14.1 Introduction 543
14.2 Vector Autoregressions 544
14.3 Problems with VAR Models in Practice 546
14.4 Unit Roots 547
14.5 Unit Root Tests 548
Dickey-Fuller Test 548
The Serial Correlation Problem 549
The Low Power of Unit Root Tests 550
The DF-GLS Test 550
What are the Null and Alternative Hypotheses in Unit Root Tests? 550
Tests with Stationarity as Null 552
Confirmatory Analysis 553
Panel Data Unit Root Tests 554
Structural Change and Unit Roots 555
14.6 Cointegration 556
14.7 The Cointegrating Regression 557
14.8 Vector Autoregressions and Cointegration 560
14.9 Cointegration and Error Correction Models 564
14.10 Tests for Cointegration 565
14.11 Cointegration and Testing of the REH and MEH 566
14.12 A Summary Assessment of Cointegration 568
Summary 569
Exercises 570
CONTENTS xv

15 Panel Data Analysis 573

What is in this Chapter? 573
15.1 Introduction 573
15.2 The LSDV or Fixed Effects Model 574
15.3 The Random Effects Model 575
15.4 Fixed Effects Versus Random Effects 578
Hausman Test 578
Breusch and Pagan Test 579
15.5 The SUR Model 579
15.6 Dynamic Panel Data Models 580
15.7 The Random Coefficient Model 581
Summary 583

16 Large-Sample Theory 585

What is in this Chapter? 585
16.1 The Maximum Likelihood Method 585
16.2 Methods of Solving the Likelihood Equations 586
16.3 The Cramer-Rao Lower Bound 588
16.4 Large-Sample Tests Based on ML 588
16.5 GIVE and GMM 589
Summary 591

17 Small-Sample Inference: Resampling Methods 593

What is in this Chapter? 593
17.1 Introduction 593
17.2 Monte Carlo Methods 594
More Efficient Monte Carlo Methods 595
Response Surfaces 595
17.3 Resampling Methods: Jackknife and Bootstrap 595
Some Illustrative Examples 597
Other Issues Relating to Bootstrap 598
17.4 Bootstrap Confidence Intervals 599
17.5 Hypothesis Testing with the Bootstrap 599
17.6 Bootstrapping Residuals Versus Bootstrapping the Data 600
17.7 NonllD Errors and Nonstationary Models 601
Heteroskedasticity and Autocorrelation 601
Unit Root Tests Based on the Bootstrap 601
Cointegration Tests 601
17.8 Miscellaneous Other Applications 602
xvi CONTENTS

Summary 602

Appendices 605
Appendix A: Data Sets 605
Appendix B: Data Sets on the Web 613
Appendix C: Computer Programs 615

Index 617