Introduction to Econometrics

Citation

Stock, J. H., & Watson, M. W. (2018). Introduction to Econometrics (4th ed.). Pearson.

Chapter Summary

The fourth edition runs to about 755 pages in five parts and nineteen chapters. Pearson describes the book's philosophy as letting applications drive the theory rather than the reverse, with real-world questions and data introducing each method, and notes that this edition adds a chapter on big data.

Chapters 1–3: Economic Questions and Data; Review of Probability; Review of Statistics

Part I, "Introduction and Review", opens by framing econometrics through economic questions and the data used to answer them, then provides two review chapters, one on probability and one on statistics. These chapters prepare the ground for regression and let a reader with rusty background refresh before the main material starts.

Chapters 4–5: Linear Regression with One Regressor; Regression with a Single Regressor: Hypothesis Tests and Confidence Intervals

Part II, "Fundamentals of Regression Analysis", begins with the simplest regression model, a single explanatory variable, and then devotes a second chapter to inference for that model through hypothesis tests and confidence intervals.

Chapters 6–7: Linear Regression with Multiple Regressors; Hypothesis Tests and Confidence Intervals in Multiple Regression

The same two-step pattern is repeated for several regressors: first the model, then the tests and confidence intervals that go with it.

Chapters 8–9: Nonlinear Regression Functions; Assessing Studies Based on Multiple Regression

Chapter 8 extends regression beyond straight-line relationships. Chapter 9 closes Part II by turning from technique to evaluation, asking how studies built on multiple regression should be assessed.

Chapters 10–14: Regression with Panel Data; Regression with a Binary Dependent Variable; Instrumental Variables Regression; Experiments and Quasi-Experiments; Prediction with Many Regressors and Big Data

Part III, "Further Topics in Regression Analysis", presents five separate extensions: data that follow the same units over time, outcomes that take only two values, instrumental variables, experimental and quasi-experimental evidence, and prediction with many regressors. The last of these is the new big-data chapter of the fourth edition.

Chapters 15–17: Introduction to Time Series Regression and Forecasting; Estimation of Dynamic Causal Effects; Additional Topics in Time Series Regression

Part IV, "Regression Analysis of Economic Time Series Data", introduces time-series regression and forecasting, then moves to the estimation of dynamic causal effects, and ends with a chapter of additional topics.

Chapters 18–19: The Theory of Linear Regression with One Regressor; The Theory of Multiple Regression

Part V, "The Economic Theory of Regression Analysis", collects the theory for the single-regressor and multiple-regressor models in two closing chapters, separate from the applied chapters that precede them.

Key Concepts

Part I: Introduction and Review (Chapters 1–3)

• Economic questions and data: the starting point of the book, which treats an empirical question and the data available to answer it as the basis for choosing a method.
• Probability and statistics review: the background in probability and statistics that regression analysis builds on.

Part II: Fundamentals of Regression Analysis (Chapters 4–9)

• Regression with one regressor: fitting a line that relates an outcome to a single explanatory variable.
• Hypothesis tests and confidence intervals: the tools for judging whether an estimated coefficient differs from a stated value and how precisely it is estimated.
• Multiple regression: holding several explanatory variables in the same model.
• Nonlinear regression functions: relationships whose effect is not a constant slope.
• Assessing studies: judging whether a regression-based study supports the conclusions drawn from it.

Part III: Further Topics in Regression Analysis (Chapters 10–14)

• Panel data: observations on the same units across several periods.
• Binary dependent variable: regression when the outcome is a yes-or-no variable.
• Instrumental variables regression: a way of estimating effects when the explanatory variable is correlated with the error.
• Experiments and quasi-experiments: evidence from randomised or naturally occurring variation.
• Prediction with many regressors: forecasting with a large set of potential predictors, the subject of the big-data chapter.

Parts IV and V: Time Series and Theory (Chapters 15–19)

• Time-series regression and forecasting: modelling and predicting variables observed over time.
• Dynamic causal effects: how a change in one variable affects another over later periods.
• Theory of linear regression: the formal development of the one-regressor and multiple-regressor models.

Critical Analysis

Strengths

• A stated applied orientation that the structure reflects. Pearson says applications should drive the theory, and the contents list is arranged accordingly: the book opens with economic questions and data, moves through regression, extensions and time series in Parts II to IV, and places the formal theory in the final part rather than at the front.
• Wide coverage in one volume. Cross-sectional regression, panel data, binary outcomes, instrumental variables, experiments and quasi-experiments, time-series regression and forecasting, and a chapter on prediction with many regressors all appear within nineteen chapters, so a single volume spans both cross-sectional and time-series methods.
• Built-in review. Two early chapters review probability and statistics, so the book does not depend entirely on a separate statistics course being fresh in the reader's mind.
• Newest material. The publisher highlights the new chapter on prediction with many regressors and big data as the main addition of the fourth edition.

Limitations

• Theory is deferred. Because the formal treatment sits in Part V, a reader who wants the derivation behind each method as it is introduced has to read the two closing chapters in parallel with the earlier ones.
• Length and selection. At about 755 pages in five parts, the book offers more than one course is likely to cover, and the contents list sets no fixed path; Part III in particular contains separate topics, so a course or reader has to choose.
• One chapter on prediction. The contents list gives prediction and big data a single chapter, so it works as an introduction to that area rather than a full treatment.

Where It Fits in the Course

Stock and Watson is the main undergraduate text in the econometrics list. Two books are gentler starting points. Angrist and Pischke's Mastering 'Metrics is described as an accessible introduction to the methods used to estimate causal effects, built around five core tools, in plain language and simple notation. A. H. Studenmund's Using Econometrics is described as teaching single-equation regression through intuition and practice, with attention to specifying, diagnosing and reporting a small empirical project. Either can be read before Stock and Watson or beside it for a different explanation of the same ideas.

Wooldridge's Introductory Econometrics: A Modern Approach is the closest alternative as a main textbook; its listing describes regression, causal reasoning, hypothesis testing, panel data, instrumental variables and empirical research. Angrist and Pischke's Mostly Harmless Econometrics is a compact guide to the tools applied microeconomists use for causal questions, among them instrumental variables, differences-in-differences and regression discontinuity, and it is a natural follow-up to Chapters 12 and 13. Newbold's Statistics for Business and Economics is a general statistics text that can support the review in Chapters 2 and 3.

How to Use This Book

The book is aimed at students who have completed introductory economics and a first statistics course; Pearson positions it for courses in introductory econometrics. Chapters 2 and 3 offer a refresher, so a reader who studied statistics some time ago can use them as a check before moving on.

A practical plan:

• Core: Chapters 1 to 9 in order. They establish the regression model, inference, multiple regression, nonlinear specifications and the habit of assessing a study critically. Everything later rests on them.
• Choose by interest: In Part III, take instrumental variables and experiments and quasi-experiments (Chapters 12 and 13) if the focus is causal evidence in microeconomics, panel data (Chapter 10) for repeated observations on firms, people or countries, and the binary-outcome chapter (Chapter 11) for yes-or-no outcomes. Chapter 14 suits readers interested in prediction and large data sets.
• Time series: Part IV for macroeconomics, finance and forecasting courses.
• Theory: Part V for readers heading into further econometrics, read alongside Chapters 4 to 7 if derivations are wanted early.

Alongside the text, estimate the models on real data in statistical software, since the book is organised around empirical questions, and work whatever exercises the course sets. Pearson sells the book on its own and with MyLab Economics, an online practice platform, so the choice between formats depends on whether a course requires it.